Awqaf New Zealand

Artificial intelligence (AI) is entering the international Awqaf, foundation and charity sectors at different speeds. Surveys of nonprofit professionals and organisations in the United States, Canada, France and Germany show substantial use, but formal governance and staff preparedness often lag behind it. Awqaf-specific discussions in Indonesia, Saudi Arabia, Kuwait, Qatar, the United Arab Emirates, Malaysia, Singapore, Türkiye, Egypt and Africa now address public knowledge tools, data infrastructure, service delivery, religious authority and institutional readiness. The report widens the comparison beyond the Awqaf sector to university and foundation endowments in North America, Europe and Asia, humanitarian agencies that hold the personal data of vulnerable people, multilateral development banks led by the Islamic Development Bank, and the nonprofit programmes of the main AI providers. Sources were searched in eighteen languages. Across all of these, the public evidence shows AI being used to support people (document search, grant administration, translation, duplicate checks, service enquiries) far more than to make entrusted decisions, and the institutions with the clearest disclosures state explicitly that final authority stays with people. It then describes Awqaf New Zealand's 2026 "Year of AI", in which a small volunteer institution used ChatGPT, Claude, Gemini and Canva for multilingual research, publication, archiving, correspondence and educational media. Its central case examines the AI-assisted comparison of Qurbani service terms across jurisdictions, which exposed gaps in the sampled materials and informed a proposed nine-interest, three-layer service framework. The report offers a proportional governance model for small institutions in six gates: classify the task, authorise tools and data, verify primary sources, require qualified human decisions, release with disclosure, and monitor and correct. The model is proposed for testing, not claimed as a Shariah ruling, legal standard or audited best practice. The report is published from the angle of a small Awqaf institution, as a contribution to a worldwide discussion. The overriding lesson is that AI does not replace Awqaf professionals and volunteers; it can extend their capacity while increasing their responsibility for verification, restraint and ethical stewardship.
Keywords: Awqaf; artificial intelligence; Waqf; charitable foundations; nonprofit governance; volunteer capacity; amanah; multilingual research; Qurbani; Shariah; privacy; endowments; humanitarian data protection; development banks; Islamic Development Bank; replication.
Central proposition: AI can expand the effective capacity of small, volunteer-led Awqaf institutions. Its contribution becomes legitimate and durable only when the institution retains human authority, verifies sources, protects people and their information, distinguishes proposals from results, and remains answerable for what it publishes or does.
An adoption and governance gap. The 2026 NTEN and Bridgespan survey received 917 responses from nonprofit staff and executives; 38.56% of the 404 executive respondents said written guidance for safe and responsible AI use was already in place. A 2026 Canadian study reported that 80% of nonprofits used AI in some form, but only 10% had a formal AI policy and 64% had none. France générosités reported that 86% of surveyed professionals used AI in their work, while 47% did so outside a framework defined by their organisation. In Germany, 41% of 450 surveyed foundations used AI and a further 16% planned to. These are different populations and methods, not comparable national rankings. [1-5]
Awqaf discussions are already practical. Badan Wakaf Indonesia publishes a public AI assistant for Waqf literacy; Saudi Awqaf sector guidance treats AI and databases as infrastructure; Malaysian institutions are building AI into strategic plans; Singapore's MUIS is adopting generative AI cautiously, with its scholars keeping oversight of final religious rulings; Qatar's endowments directorate has agreed with a fintech company to develop a pilot AI platform for Shariah-compliant analysis of its endowment share investments; a 2025 seminar led by the Kuwait Awqaf Public Foundation recommended using AI in endowment management under Shariah and technical oversight; and the 12th Waqf Fiqh Issues Forum has placed digital endowments on its agenda. The debate also includes principled criticism: one Kuwaiti scholar argues that the philosophy of digital systems sits in tension with the objectives of Waqf. An announced tool, a strategic aspiration, a recommendation and an independently evaluated deployment are separate kinds of evidence. [13-18,22-23,39-41,44-48,52-57]
Endowments and foundations use AI to support people, not to replace their decisions. The University of California's investment office uses AI to monitor risk and detect anomalies, stating that "technology doesn't replace judgment". Rice University calls AI "a tireless (and sometimes overconfident) research assistant". Foundations in Finland, Norway, Denmark and Switzerland use AI in grant administration while stating that people make every award, and Finland's Kone Foundation declines to use AI in grant evaluation at all. No institution reviewed was found to let AI make endowment investment or property decisions on its own. [58-73]
Humanitarian agencies show what is at stake with personal data. WFP uses AI to flag possible duplicate aid registrations, but "the tool doesn't decide who gets assistance". The Norwegian Refugee Council is procuring a controlled AI tool because staff were already turning to consumer AI. The 2021 Rohingya data controversy shows the cost of weak consent. For UK direct marketing that uses religious belief, even when inferred, the regulator advises obtaining explicit consent, because other conditions for such data are unlikely to apply. [74-79,91]
Development banks are adopting AI with human checks, but not yet in Waqf decisions. The Islamic Development Bank uses an AI-enabled advisory agent in development effectiveness reporting and is planning a unified, governed AI platform; its Institute runs an Islamic finance AI assistant whose answers are reviewed by experts. No evidence was found that its Awqaf Properties Investment Fund uses AI to value or select Waqf property: feasibility studies are reviewed and challenged by its team. The World Bank and AIIB disclose AI use and state that people verified the content, and the Asian Development Bank builds digital risk assessments into each stage of the projects it finances. [92-105]
Public trust can move in the opposite direction from productivity. In a 2026 UK survey, 58% opposed charity videos generated with AI and 69% opposed fictional AI-generated images or stories of the people a charity helps. These are British attitudes, not those of Qurbani offerors, but they warn against synthetic media that could be mistaken for evidence about real beneficiaries. [6]
The Awqaf NZ experience. AI helped a volunteer institution publish multilingual research, rebuild its website in six languages, index fifteen years of archives and manage correspondence. The multilingual review of Qurbani terms was especially instructive: it turned scattered examples into a structured discussion of rights, responsibilities and missing provisions. It does not establish that the Qurbani industry lacks applicable law. [25-27]
A practical data-governance lesson. Google's no-cost Workspace for Nonprofits edition includes the Gemini app with enterprise-grade data protections, but tools such as Google Vault and Data Loss Prevention are listed only for paid editions. An institution must still govern data classes, sharing, retention and each product's terms; a provider commitment is not a completed compliance audit. [28-32]
The international conversation about AI has changed quickly. Some organisations want to move from experimentation to routine operations; others see serious risks to privacy, fairness, professional competence and trust. For Awqaf institutions the stakes are distinctive. An endowment can involve assets dedicated to particular purposes, a founder's stipulations, religious and legal duties, trustees or nuzzar, beneficiaries, specialist service providers and intergenerational stewardship. A faster process is not a better one if it distorts a Shariah distinction or obscures who remains accountable. [8,12,19,24]
Awqaf New Zealand approaches the question from an unusual position. It is a New Zealand charitable trust founded in 2011, whose research was developed through volunteer effort rather than a funded research department, and long before the current wave of generative AI. The change in 2026 was therefore a change in capacity to examine and communicate work already in progress, not a claim that software created the programme or replaced its human originators. [27]
This report grew out of a November 2026 discussion prompted by an essay by Dr Hisham Dafterdar, a Trustee of Awqaf New Zealand and Chairman of Awkaf Australia Ltd, which he posted publicly in the Global Awqaf Forum WhatsApp group. He asked how Awqaf bodies can use AI without replacing spiritual purpose, accountable human judgement and stakeholder trust with an impersonal logic of automation. His concerns about sensitive information, fairness and fiduciary duty provide an important counterweight to the experience of a volunteer team finding new research capacity through AI. [24]
Rather than reply briefly, Awqaf NZ chose to examine international evidence in several languages and to set out its own experience in a report open to public discussion. Dr Dafterdar's essay is a cited perspective and the prompt for this work; it is not co-authorship or endorsement, and the conclusions here should not be attributed to him or to a Board resolution.
The report pursues four objectives: (i) map significant international positions, including criticism and caution; (ii) distinguish implementation evidence from aspiration and commentary; (iii) explain Awqaf NZ's practical experience and the safeguards that guided it; and (iv) offer a testable, proportionate governance method for volunteer-led institutions.
Awqaf refers to the institutional field of Islamic endowments, and Waqf to an individual endowment or the legal institution, as context requires. Foundation refers to the legal entity described in a source; not every foundation is a Waqf, and not every Waqf has the legal treatment of a foundation. AI includes generative language models and other algorithmic systems; a document assistant, a statistical classifier and an autonomous transaction system carry very different risks.
This is a targeted multilingual narrative scoping review supplemented by an institutional case account. It is not a systematic review, a census of Awqaf institutions, a complete legal comparison, or an experiment measuring the effects of AI. Sources were identified through public institutional and academic searches in English, Arabic, French, German, Indonesian, Malay and Turkish for the Awqaf and Qurbani material. Later sections extended the search to Persian, Urdu and Bengali (national Awqaf authorities), Dutch, Danish, Norwegian, Swedish, Finnish, Japanese and Korean (foundations), and Spanish (development banks): eighteen languages in all. The searches were run with Gemini and ChatGPT, and the results were checked against the original sources with Claude. A search that found nothing in a language is not proof that nothing exists. Emphasis was on material available by 21 November 2026.
Sources were prioritised in this order: original reports with described samples or methods; regulator and organisational guidance; original Awqaf institutional tool or programme pages; scholarly work; and, where necessary, reputable news agencies reporting named official announcements. Secondary reports were not treated as proof that a promised project had achieved its outcomes. Survey percentages refer to different denominators and are grouped for qualitative comparison only. Non-English findings are paraphrased in English, with original titles kept in the references.
| Evidence category | What it can support | What it cannot support by itself |
|---|---|---|
| Survey with stated sample and method | Reported prevalence or attitudes among its respondents | A universal comparison or causal impact |
| Official Awqaf programme or tool page | A publicly announced or visible capability | Accuracy, effectiveness, adoption or assurance |
| Conference theme or policy announcement | A live institutional or scholarly discussion | Operational implementation |
| Legal, regulatory or fatwa text | The stated position of its issuing authority | Automatic application in every country or case |
| Awqaf NZ record or publication | Work reported and its stated status | Independently audited benefits or measured savings |
| AI-generated comparison or synthesis | A useful hypothesis, index or draft | A verified primary source or authoritative decision |
A second distinction runs through the whole report: a written intention, work in progress, formal approval, operating service and verified outcome are different statuses. They should not be flattened into one category labelled "innovation". The rule is applied to external sources and to Awqaf NZ's own work alike.
The following studies offer the strongest broadly relevant quantitative evidence located for this review. They describe their own respondents, not the global Awqaf sector.
| Study, year and population | Finding on use | Finding on governance | Limitation |
|---|---|---|---|
| NTEN and Bridgespan, 2026: 917 nonprofit professionals, mainly in the US | Adoption and expectations differ between executives and frontline staff | 38.56% of 404 executives reported written guidance in place | Individuals, not a census of organisations [1] |
| Imagine Canada and partners, 2026: national survey of nonprofit organisations | 80% reported AI use in one or more activities | 10% had a formal AI policy; 64% had none | Canadian sector only [2] |
| France générosités, June 2026: 375 professionals from 199 organisations | 86% used AI in their work | 47% used it outside a framework defined by their organisation | Individual respondents, not organisation-wide uptake [3] |
| France générosités and Don en Confiance, March 2026: practical guide | Not a survey | Recommends assigning responsibilities and adopting an ethical charter or AI use policy | Guidance, not measurement [4] |
| Bundesverband Deutscher Stiftungen, 2026: 450 foundations | 41% used AI; 16% planned to | Practical use often runs ahead of formal rules | Civil-law foundations, not Islamic Awqaf [5] |
The recurring pattern is that individual enthusiasm and experimentation outrun organisational rules, shared training and accountability. An organisation can use AI every day without having answered basic questions: who is authorised to use it, what data are allowed, what must be checked, and who approves the result? [1-5]
The German study also finds that AI use correlates with organisational size and staffing. That suggests a risk of a new capability divide for volunteer-led bodies, even when AI tools are widely available. Awqaf NZ's experience points to one response: pooled skills, reusable workflows and shared verification, provided the safeguards travel with the methods. [5]
Usage statistics are easier to collect than public-benefit outcomes. A charity may produce more messages, but that does not prove recipients were better served; an automated screening process may be faster, but that does not prove it was fair. The Center for Effective Philanthropy found that 90% of surveyed foundation leaders said their foundations offered no funding or nonmonetary support to grantees for AI use, which leaves open whether AI capability helps grantees fulfil their missions at all. [7]
Regulators reinforce the point. The Australian Charities and Not-for-profits Commission advises charities to manage the risks specific to their work, including data quality, and reminds responsible people of their duties under the Governance Standards, noting that Australia has no formal regulatory framework for AI. The Charity Commission for England and Wales, in its Charity Sector Risk Assessment 2026, identifies AI-assisted fraudulent registration and grant applications among emerging technology risks and expects trustees to show active oversight. [8-9]
A YouGov study published on 19 November 2026 surveyed 2,097 adults in Great Britain in August 2026. Majorities opposed fictional AI-generated stories or images of beneficiaries (69%), donation appeals personalised by AI from giving history (59%) and AI-generated charity videos (58%). Summarising internal reports was the most accepted use (42% acceptable, 32% not), yet no tested use reached majority acceptance. These are UK adult attitudes, not a study of Qurbani offerors. [6]
The implication for Awqaf NZ is not "do not make cartoons". A clearly fictional Animal Family explaining stewardship is materially different from a generated photograph implying that a real family received a service. The proposed rule separates educational imagination from operational evidence: show cartoons as cartoons, state factual promises in ordinary text, and evidence service fulfilment with authentic records, never generated scenes.
Large university endowments share some of the Awqaf sector's defining concerns: capital intended to last indefinitely, spending rules and duties to future generations. They are secular investment pools, so the comparison concerns governance, not religious purpose. This review distinguished AI that an endowment office uses from AI companies it invests in; only the former is reported here.
Public disclosure is sparse. No internal use of AI by the investment offices of Harvard, Yale, Stanford, MIT, Princeton or the Wellcome Trust could be verified from public sources. A 2024 Harvard master's student project built a prototype to extract terms from limited partnership agreements with Harvard Management Company staff, but its authors described deployment as hoped for, not achieved. [65]
Where disclosure exists, it describes support work under human authority. The University of California's investment office states that AI "informs how we monitor risk, analyze performance, and even detect anomalies in billion-dollar pension flows", and that "technology doesn't replace judgment, it enhances it". Rice University's 2025 endowment report describes AI as "a tireless (and sometimes overconfident) research assistant", introduced through a "crawl-walk-run" approach in a secure "walled garden" so that "innovation doesn't outpace oversight". Global Endowment Management, which invests for endowments and foundations, uses AI to speed data gathering and meeting preparation but has "deliberately stopped short of delegating investment decision making to AI". [58-60]
Others are preparing carefully. Cambridge University's investment office first migrated its IT systems and gained Cyber Essentials security accreditation, and only then planned its "first, measured steps" with AI tools in 2026. Oxford University's endowment office reported, through the consultancy that built it, a tool that lets analysts query quarterly manager reports with answers linked to their sources; it was described in late 2024 as "being tested in daily operations". The University of Texas investment company's board was briefed on staff AI upskilling in 2026, and the manager of the Gates Foundation Trust's endowment is recruiting to expand generative AI in its legal team. [61-64]
A search in German, French, Dutch, Danish, Norwegian, Swedish, Finnish, Japanese and Korean found the strongest evidence of foundations using AI in their own administration, mostly in handling grant applications. [66-73]
| Foundation | What it does with AI | Stated limit or status |
|---|---|---|
| Finnish Cultural Foundation (Suomen Kulttuurirahasto) | AI functions in Microsoft Azure OpenAI help organise and summarise grant applications, since autumn 2025 | "AI does not make decisions about grant recipients"; application data are not used to train models |
| Sparebankstiftelsen DNB, Norway | AI assists in processing applications up to NOK 50,000 and drafts grant summaries | "All applications are still assessed by the administration and decided by the board"; texts are reviewed before publication |
| SpareBank 1 Stiftelsen BV, Norway | AI assists in assessing applications; an applicant assistant ("Iver") | Discloses that application content is shared with OpenAI; conversations kept up to 60 days |
| Villum Fonden, Denmark | AI may assist in preparing and processing applications | "The foundation does not use AI for the final decision" on awards; all AI output is checked by people |
| Swiss National Science Foundation | Testing machine learning to suggest suitable reviewers | Reviewers form their own opinion first; AI "can support their work, not replace it" (rules of 20 August 2026) |
| VolkswagenStiftung, Germany | Internal AI for translation and text drafting | Under an internal AI guideline on data protection and integrity |
| Dutch cultural funds | Controlled pilots in grant administration | Advisory committee members may not use generative AI in substantive assessment |
| Kone Foundation, Finland | None in grant decisions | "Does not use AI to evaluate applications or decide funding" |
Two points stand out. First, the same technology is governed in opposite ways by respected institutions: some use AI to support grant handling, others refuse it in assessment, and both positions are legitimate choices. Second, every disclosure that permits AI also states where human authority begins.
Humanitarian agencies hold the personal data of people who often cannot refuse to give it, which makes them the most relevant comparison for Qurbani recipients. The International Committee of the Red Cross's Handbook on Data Protection in Humanitarian Action (third edition, 2024) includes a chapter on AI and treats consent with caution where aid depends on registration. [74]
WFP's Enterprise Deduplication Service compares names and photographs to flag possible duplicate registrations for food assistance. It is deployed in Mali and piloted in seven other countries, and WFP reports savings of more than US$431,000 in six months of 2025. A staff member reviews every flag: "the tool doesn't decide who gets assistance". Each new context requires a privacy assessment before rollout. [75]
The Norwegian Refugee Council's April 2026 tender for an AI tool is unusually candid: "increasingly NRC staff are turning to consumer AI tools". Its response was not prohibition but a controlled tool for more than 500 staff, requiring that prompts "remain within the NRC tenant and are not used in general training", with GDPR and EU AI Act compliance. NRC's privacy policy draws a distinction directly relevant to Qurbani distribution: it may decline assistance only where a person objects to processing strictly necessary for eligibility, and must continue to serve people who object to optional uses of their data. [76-77]
The cost of weak consent is documented. In 2021 Human Rights Watch reported that Rohingya refugees registered by UNHCR in Bangladesh had their data passed by the Bangladeshi government to Myanmar, and that most of those interviewed did not understand this would happen; UNHCR disputed parts of the account. This was a data-governance failure, not an AI failure, but AI makes such data easier to match and share. [78]
The major AI providers now offer discounted nonprofit plans and governance guidance. The table records facts from each provider's own pages as at November 2026; prices change and should be rechecked.
| Provider | Nonprofit offer (as published) | Governance resource | Data training default (business plans) |
|---|---|---|---|
| Gemini app included in the no-cost Workspace for Nonprofits edition; paid editions discounted | Google.org AI Readiness Playbook for Funders (2026) | Not used to train models | |
| Microsoft | Discounted Microsoft 365 and Copilot plans | AI Governance Framework for Nonprofits (2024) | Organisational data not used to train models |
| Anthropic | Claude Team at US$8 per user per month; about US$3 in low and middle-income countries; up to 75% off Enterprise | AI Fluency for Nonprofits course | Not used to train models by default |
| OpenAI | ChatGPT Business at US$8 per user per month (annual); up to 75% off Enterprise | Nonprofit resources in OpenAI Academy | Not used to train models |
Every programme leaves the same responsibilities with the charity: classifying data before it is entered, checking outputs, budgeting for ongoing costs, and deciding what AI may and may not do. A discount does not transfer any of these duties. [28-29,80-82]
Disclosure: this report was written largely with Claude (Anthropic) and ChatGPT (OpenAI), with reviews by Gemini (Google), and Awqaf NZ uses Google Workspace for Nonprofits. Each provider is therefore described in the same format, using only its own published information, and no ranking or recommendation is made.
Multilateral development banks matter to the Awqaf sector for two reasons. They finance Waqf property and Islamic social finance directly, and they have begun to publish how they use AI themselves. Their evidence needs to be read on two separate layers: the AI a bank uses inside its own offices, and the safeguards that apply to people affected by the projects it finances. A sound internal AI policy does not show that a borrower's beneficiary registry, biometric system or contractor meets the same standard. [92-105]
The Islamic Development Bank Group. The IsDB is the most relevant bank for Awqaf institutions, because its Awqaf Properties Investment Fund (APIF) has financed income-generating Waqf property in member and non-member countries since 2001. Four kinds of evidence are now public, and they should not be merged:
What was not found matters as much. No public source shows APIF using AI to value Waqf property, interpret deeds, select projects or allocate investment. APIF's own guidance says a feasibility study, usually prepared by an independent consultant, "is reviewed and challenged by the APIF team during the appraisal phase". The Fund also asks applicants for Waqf deeds, organisational records and trustee details, which shows why any future document-reading AI at a Waqf financier would need a clear privacy assessment first. Several Awqaf-related programmes in the Institute's 2025 report, such as Awqaf Free Zones, appear beside IFAA but should not be presumed to use AI. [95,98]
Other development banks. The wider pattern is the same: AI is used to help staff, outputs are checked by people, and use is disclosed.
| Bank | Internal AI use found | Oversight and disclosure | Status in this report |
|---|---|---|---|
| World Bank | Internal mAI interface using ChatGPT 5 and Copilot | World Development Report 2026 states that "all analytical judgments, policy interpretations, and conclusions were developed, reviewed, and verified by the author team" | Verified |
| Asian Development Bank | AI-assisted document review planned in an evaluation of the Asian Development Fund | Digital ADB 2026-2030 (July 2026) commits to AI-specific governance and digital risk assessments "systematically embedded at key stages of the project lifecycle"; the evaluation approach paper (August 2026) states that "all AI-assisted outputs will be subject to human review, crosschecking, and triangulation" | Verified |
| Asian Infrastructure Investment Bank | AI used to help prepare a 2026 publication; its 2025 annual report states that it launched an internal AI platform giving staff "secure, institutionally governed AI capabilities" | "All content was reviewed by AIIB staff" | Disclosure verified; platform stated by AIIB (performance not independently assessed) |
| Inter-American Development Bank | IINA, an in-house assistant that searches more than 900 public documents, and Aurora, an internal generative AI multi-assistant platform | Listed among "projects, prototypes and pilots" in its Tech Report: Generative AI (March 2025) | Verified as prototypes or pilots; operating status not stated |
| EBRD | An "integrity chatbot" for compliance staff, reported as in development | Its 2024 Environmental and Social Policy is reported to cover data protection and privacy risks in client activities | Claimed, not independently confirmed |
The most useful lesson for Awqaf institutions is ADB's: it treats AI risk as part of project appraisal, at each stage of a project's life, rather than as a separate technology question. The World Bank and AIIB disclosures show a second habit worth copying: saying plainly that AI was used and that people checked the result. [99-105]
A Gemini search also reported that Indonesia's national zakat agency, BAZNAS, uses an AI-supported system for zakat management, citing an article in the agency's research journal. This report could not confirm the system's scope or status, and it is recorded as a claim only. [106]
Questions for any AI-supported Waqf or Zakat project financed by a development bank. Before such a project goes ahead, an Awqaf institution should be able to answer: who controls the personal data of founders, beneficiaries, tenants and trustees; whether a beneficiary can use a non-biometric or non-automated route; whether the provider may reuse the records; how a wrong fraud or eligibility flag is explained and corrected; and which Shariah authority approves eligibility rules and asset use. The public sources reviewed do not answer these questions for any bank. They are a checklist, not a finding of failure.
Badan Wakaf Indonesia (BWI) offers a public Literasi Wakaf AI Assistant that answers questions about types of Waqf, regulations, how to make a Waqf and how to become a professional nazhir, with a separate disclaimer page. It is a visible example of an Awqaf authority using AI for public literacy. The public page does not show how answers are reviewed, so it cannot by itself establish their completeness or suitability for legal reliance. [13]
The governance question is twofold: how can an institution make Waqf information accessible in everyday language, and how can it prevent an interactive explanation from being mistaken for a binding ruling? The question is sharper for multilingual audiences unfamiliar with technical fiqh terms.
The Saudi General Authority for Awqaf's sector-enablement material identifies technology services, databases and AI as part of the sector's technical infrastructure, alongside legal and regulatory enablement, reliable specialist information, governance, human capability and coordination. This is best read as an institutional-development agenda, not evidence that a particular AI application has delivered measured benefits. [14]
The lesson is that AI is most valuable where reliable legal records, trustworthy data and trained administrators already exist. AI cannot repair missing ownership records or reconcile contradictory stipulations simply by being sophisticated.
On 14 May 2026 the Federal Territories Islamic Religious Council (MAIWP) launched its Strategic Plan 2026 to 2030, which integrates advanced technologies including AI so that decisions are data-driven and "no asnaf is left behind" as the number of new zakat recipients grows. A strategy launch is not an audit of eligibility outcomes. [15]
On 27 July 2026, at the International Waqf Dakwah Convention in Kuala Lumpur, Malaysia's Prime Minister urged waqf institutions to embrace digitalisation and AI without losing faith, ethics and moral character; the convention resolved, among other things, to use technology and digitalisation in waqf management. In September 2026 the 12th International Conference on Mosque, Zakat and Waqf (IMAF 2026) at Universiti Islam Selangor took "Digital Transformation and Artificial Intelligence in the Management of Mosques, Zakat and Waqf" as its theme. Malaysian institutions are therefore debating both the opportunity to understand need more accurately and the ethics of how assistance decisions are made. [16-17]
Burcu Şenlikoğlu's 2025 Turkish-language article examines AI in foundation activities through legal regulation and application examples. It finds AI usable in the establishment, management and purpose-related activities of foundations, while discussing its negative aspects and the personal-data questions it raises. Its legal framing is a useful corrective to promotional language that treats an AI system as if it could itself hold a fiduciary office. It is a scholarly analysis of its own jurisdiction, not a rule for Awqaf bodies elsewhere. [18]
Egypt's Dar al-Ifta ruled in September 2025 that relying on AI applications to obtain fatwas is not permissible, because no software yet matches the qualified mufti's ability to weigh time, place and personal circumstance. A separate 2025 ruling permits AI as an aid in producing scholarly content where the author remains the main contributor, verifies the information and attributes it, but forbids presenting wholly AI-generated work as one's own. In March 2026 it allowed reliance on AI-assisted medical reports reviewed, approved and signed by a supervising physician, while rejecting reliance on unsupervised AI output. [19-21]
These rulings answer different questions and should not be merged into one edict about technology. Together they illustrate a principle every Awqaf institution can use: the permissibility and risk of assistance depend on the task and the authority it requires. A model locating a source, drafting an explanation, recommending a beneficiary or purporting to issue a fatwa are not interchangeable activities.
In February 2025 the Kuwait Awqaf Public Foundation, with the International Islamic Charity Organization and the International Forum of Development Organizations, held a seminar on digital endowments at the International Islamic University Malaysia in Kuala Lumpur. Its sessions covered the Shariah basis, legal form, governance, technical features and practical applications of digital endowments, including a paper on blockchain and AI. The participants' adopted recommendations call for using AI to analyse digital endowment data, improve investment efficiency and support strategic decisions. In the same document they call for a central body and clear regulations for digital endowments, Shariah and technical standards, Shariah and technical review committees, periodic transparency reports, and KPIs to measure developmental and financial impact. AI is therefore placed inside a governance frame rather than treated as an end in itself. These are recommendations, not adopted regulation or evidence of systems in operation. [39]
The seminar's working papers also record genuine scholarly disagreement. Dr Sara Mutla Al-Qahtani of Kuwait University argued that the philosophy of digital systems sits in tension with the Shariah objectives of Waqf: data held as an endowed asset may lack provenance and may not be under its owner's free and lawful control; digital norms tend to direct use towards material ends that disregard founders' stipulations; and absorbing Waqf into digital systems risks reducing it to a financing function. She concluded that a settled position on the governance (nazara) of digital endowments requires data-governance initiatives and ethical guidance that treat digital tools as a means, not an end. This is one scholar's argument, not a consensus and not a finding that AI is impermissible, but it is a useful corrective to purely enthusiastic accounts. From a technical perspective, Dr Omran Salem identified regulatory gaps, cybersecurity risks and low technological literacy as the main challenges, and proposed independent, periodic auditing and certification of digital Waqf platforms against security, Shariah, administrative and data-protection standards. [41]
The Foundation's 12th Waqf Fiqh Issues Forum, held every two years, has chosen two topics of direct relevance. Its first topic is the rulings on Temporary Cash Waqf, including the similarities and distinctions between Temporary Cash Waqf and Qard Hasan. Its second is the Shariah basis of digital endowments: their definition, legitimacy and forms, Shariah audit, who may act as nazir and under what conditions, changes to a founder's stipulations, the use of income, and practical applications. The topic is framed around electronic platforms and emerging technologies. The Forum's discussion of digital Awqaf is relevant to future questions about AI-assisted administration, but its published research brief does not expressly determine those questions. As far as Awqaf NZ is aware at the time of writing, the Forum has not yet convened or issued resolutions, so its position should not be anticipated. [40]
Disclosure: Awqaf New Zealand has a direct interest in the Forum's first topic. Its Temporary Cash Waqf model depends on the distinction from Qard Hasan, and the author has followed and commented on research prepared for that topic (Section 7.1).
Singapore is widely regarded as a model of professional Awqaf administration, and it shows how established governance can frame AI adoption. Under the Administration of Muslim Law Act (AMLA), ownership of all Wakaf assets is vested in the Islamic Religious Council of Singapore (MUIS). The assets are managed by its wholly owned subsidiary, Warees Investments, and by mutawallis approved by the MUIS Council, in accordance with each founder's intentions; mutawallis must also submit audited financial statements. Since 2012 the Wakaf Revitalisation Scheme, administered by Warees, has identified properties that can be enhanced or redeveloped for better returns to beneficiaries, while asset migration (istibdal) rests on guidance given by Singapore's Fatwa Committee in October 1985. MUIS's Chief Executive described this work at the 2026 Wakaf Disbursement as "protecting an amanah", announcing S$8.48 million in Wakaf income to be disbursed that year. [44-46]
MUIS's approach to AI is deliberately cautious. In a parliamentary reply of 26 February 2026, MUIS said it adopts generative AI judiciously and with safeguards, warned of the risks when AI is relied upon as the primary source of religious guidance, and stated that qualified asatizah "continue to maintain oversight on final religious rulings". IftaSG, an initiative of MUIS's Fatwa Lab, uses prompt engineering and curated datasets to tailor a large language model for fatwa research suited to Singapore; the Fatwa Lab's first research series included a study on optimising fatwa research with AI. MUIS also reports stepped-up AI and digital-literacy training for mosque and madrasah staff and volunteers. In August 2026 the Acting Minister-in-charge of Muslim Affairs highlighted IftaSG and a Halal Ingredients Verification Bot as examples of AI innovation; the release does not say whether either is launched, piloted or planned. [47-49]
Two further items are relevant but carry less weight. First, a publicly posted LinkedIn account by an AI Singapore trainer reports an AI for Essential Digital Skills (AI4EDS) train-the-trainer workshop held in 2026 for the Board of Directors and senior management of Warees Investments, covering productivity, decision-making, knowledge management and stakeholder engagement. This is the trainer's own claim; it could not be independently verified for this report, and it describes capacity-building, not AI deployed in asset management. Second, in August 2026 Warees visited Indonesia's Dompet Dhuafa to exchange experience on Waqf real-estate portfolios, digital channels and donor relationship management, with a view to developing digital Waqf services in Singapore; the published account does not mention AI. [50-51]
The public sources reviewed do not show AI being used for Wakaf property valuation, investment allocation or asset decisions at Warees. Singapore's lesson for this report is about sequence: clear legal ownership, professional asset management and religious oversight came first, and AI is now being introduced to support specialist work without displacing the people accountable for it.
Qatar provides the clearest planned example found in this review of AI being applied to the management of endowment assets themselves. On 9 June 2026 the General Directorate of Endowments at the Ministry of Endowments and Islamic Affairs signed a memorandum of understanding with Wahed MENA, a Qatar-based Islamic fintech company, to develop and launch a pilot AI platform that classifies and analyses shares listed on the Qatar Stock Exchange for Shariah compliance and supports endowment investment decisions. The announced features include dividend screening, tracking Shariah compliance over time, investment risk analysis and an intelligent assistant to support internal Shariah-based analysis. It is an agreement to develop a pilot, not an operating or evaluated system, and the announcement does not describe how its outputs will be reviewed or approved. [52]
The Ministry's wider programme is substantial. In December 2025 it announced a strategic partnership with Google Cloud, including an "Awqaf AI Factory" for data processing and model training on its archives, and migrated its Islamweb platform to Google Cloud. It launched a "Live Fatwas" AI service on Islamweb in September 2025 and the FatwaTok application in March 2026. At a forum the Ministry co-organised in July 2026, panellists stressed that "human oversight remains the kernel source in rendering Fatwas and creating religious content". [53-55]
In the United Arab Emirates, the federal General Authority of Islamic Affairs and Endowments announced in October 2025 that more than 5,000 employees would be trained in AI, in collaboration with unnamed technology companies. In October 2026 Awqaf Abu Dhabi presented a photoreal AI agent as a single point of contact for its endowment and minors' funds services, alongside a unified digital platform and a smart application planned in two phases, minors' funds first and endowments second. These are announcements of training and service channels, not evidence of AI in endowment asset decisions. [56-57]
A search of national Awqaf authorities in Arabic, Persian, Turkish, Urdu, Malay, Indonesian, French and Bengali found that most have invested in digital records, and few have yet applied AI to them. [83-90]
| Authority | What the sources show | Status |
|---|---|---|
| Egypt, Ministry of Awqaf | An AI Centre opened by the Prime Minister on 2 November 2026 to produce religious educational content | Operating; religious outreach, not property |
| Egypt, Egyptian Awqaf Authority | Electronic archive of Waqf deeds and GIS maps of endowed land, under a 2016 presidential decree | Digital foundation; no AI reported |
| Iran, Awqaf and Charity Affairs Organisation | Big-data valuation software that suggests a price range has been developed and is intended as a pricing reference in judiciary systems; officials want to extend AI use in property appraisal | Proposed; routine operation not confirmed |
| Türkiye, General Directorate of Foundations | Digital deed archive and property inventory systems | Digital foundation; no AI reported |
| Morocco, Ministry of Habous | 40,000 pages of Habous documents digitised in 2024 | Digital foundation; no AI reported |
| Jordan, Ministry of Awqaf | GIS records of Waqf property; electronic services for Waqf investors from September 2026 | Digital foundation; valuation by accredited valuers |
| Oman, Ministry of Endowments | Online management of Waqf asset income and expenses; staff AI training | Digital service; training |
| Saudi Arabia, General Authority for Awqaf | Procurement notice for AI analysis of contact-centre calls; training course on AI in the Waqf sector | Procurement (not independently checked) |
| Algeria, Ministry of Religious Affairs | National Waqf property database planned; AI added to staff training | Planned; training |
| Sudan, Iraq, Pakistan, Brunei | Digital registration services in some cases; no AI in Waqf management found | Not verified |
| Tunisia | Its Habous system was abolished in 1956 and 1957, so there is no comparable authority | Not comparable |
The pattern matters for small institutions. Large authorities have spent years building digital deed and property records, and those records are what would make AI search or valuation possible and checkable. Iran's proposed AI-assisted valuation is the only national case found of AI being prepared for Waqf property valuation, not an established decision system, and it raises the questions any such system must answer before use: who chooses the comparable properties, whether a licensed valuer signs off, how founders' unusual stipulations are respected, and how tenants can appeal.
In 2026 Awqaf Africa published a governance-first AI readiness framework for waqf, zakat and Islamic social-finance institutions (GF-AIRISF), with eight dimensions linking institutional accountability, Shariah and maqasid, data, security, people, operations, law and assurance. A second paper asks when AI may assist, recommend, execute under supervision, or must leave a decision to humans. Both are useful evidence of the sector's conceptual debate and describe themselves as frameworks for further validation, not adopted standards. [22-23]
Taken together, these sources show a growing conversation across several regions, but few openly documented and independently evaluated Awqaf AI deployments. That is a provisional observation from this review, not a worldwide census.
The strongest opportunity for a small Awqaf institution is capacity amplification. Work that once required several languages, long manual searches and repeated formatting can be accelerated, leaving human experts to concentrate on interpretation and judgement. The benefit is greatest where questions are clearly formulated and the underlying records can be inspected.
Multilingual discovery and comparison. AI can propose search terms, organise documents and surface missing categories across Arabic, English, French, German, Indonesian, Malay and Turkish material. It cannot by itself establish that a translated legal concept carries the same meaning in each jurisdiction.
Institutional archives and memory. Indexing and summarising historical material helps a volunteer institution recover why earlier decisions were made and avoid mistaking an abandoned proposal for current policy, provided summaries link back to dated originals.
Accessible educational communication. Plain-language explanations, infographics, translations, audio and recognisably fictional characters lower barriers for people unfamiliar with Awqaf terms. Views and page counts do not by themselves measure understanding or trust.
Service design and quality assurance. AI can turn a broad service into a checklist of responsibilities, exceptions and failure points, and highlight inconsistencies in draft terms. People must still decide what is legally required, what is contractually promised and what is merely good practice.
Planning and administration. Approved tools can support meeting preparation, research indexes, work registers and other repetitive tasks. Savings should be measured net of review, correction, training, licensing and security costs; free software does not make work costless.
Shared capability across institutions. Small Awqaf bodies can share terminology, disclosure clauses, research protocols and governance questions. Sharing a method is not sharing legal permissions, personal data, contracts or approvals.
Language models can fabricate references, misattribute quotations or collapse distinctions between similar concepts, which is especially damaging for contracts and fiqh. A generated claim that Temporary Cash Waqf and Qard Hasan are interchangeable would bypass the very distinction Awqaf NZ has spent years researching. A second model agreeing with the first does not validate the claim: cross-model disagreement can show what to investigate, but agreement is not proof. Primary texts, independent review and qualified scholars must stay in the chain of evidence. [19-21]
New Zealand's Privacy Commissioner expects organisations to involve senior leaders and privacy officers before adopting generative AI, to assess privacy impacts, and to avoid entering confidential or personal information into tools that may retain or disclose it. The Privacy Act 2020 applies to nonprofits as it does to other agencies. Records of religious giving or beneficiary eligibility can carry particularly sensitive implications. [10]
A proportionate rule is to use public or de-identified data for general AI research, keep identity-bearing files under approved access controls, and assess providers and law before any sensitive workflow. A seemingly harmless spreadsheet can identify people through combinations of location, age, payment or household details. CommunityNet Aotearoa's guidance for community organisations recommends an explicit "do not use AI" list, including eligibility decisions and client case notes. [11]
Donors need separate protection. Donor records reveal giving history, wealth and, for a Waqf or Qurbani service, religious practice. The UK Fundraising Regulator's AI guidance (December 2025) makes charities accountable for AI used in fundraising, requires a risk assessment before any AI tool is used, requires people to check AI content for "accuracy, fairness and legality", and advises: "do not feed information not already in the public domain into publicly available or open access AI". For UK direct marketing involving information about religious beliefs, including inferred beliefs, the Information Commissioner's Office advises obtaining explicit consent, because other special-category processing conditions are unlikely to apply; the applicable lawful basis and marketing rules must also be assessed. An Awqaf institution should therefore never infer a person's faith from a name, a location or a past Qurbani order in order to target appeals. [79,91]
Recipients and donors should not be pooled. Recipient data serve eligibility, protection and fair distribution; donor data serve receipts, stewardship and lawful fundraising. Following the NRC principle in Section 3.6, recipients should be told which information is strictly necessary for a service and which is optional, and should never lose assistance for refusing an optional use. [77]
An algorithm that ranks potential beneficiaries can amplify biased records or treat unusual circumstances as errors. For high-impact decisions an Awqaf body should explain its criteria, offer human review, test for unequal outcomes and keep the authority to correct the record, consistent with regulator guidance and UNESCO's ethics recommendation. [8,12]
Synthetic testimonials, fabricated beneficiary images and unlabelled generated "evidence" threaten the relationship between an Awqaf institution and its public. The UK findings suggest that people may accept internal summarisation while rejecting synthetic content used to seek support. For faith-based educational cartoons, the answer is transparent fiction, not deceptive realism. [6]
AI can reduce drafting effort while increasing what must be checked; if output grows faster than verification capacity, the bottleneck simply moves downstream. Licences, limits, interfaces and data terms also change. An Awqaf body should keep exportable source files, a human-readable knowledge register, fallback methods and more than one person able to run essential workflows.
Computing has energy and material costs. An institution claiming stewardship of resources should ask, in line with UNESCO's principle of proportionality, whether a simpler method would be adequate before adopting a more resource-intensive one. [12]
Awqaf institutions may be subject to charity, privacy, consumer, intellectual-property, payment, Awqaf, food-safety and other law even without a dedicated AI statute. Claims that a sector is "unregulated" need careful jurisdictional qualification, especially where a study samples provider terms rather than statutes.
Awqaf New Zealand was founded in February 2011 and is a registered New Zealand charity (CC46694). Its Smart Waqf research, Shariah consultations, contract drafting and international exchange were developed by volunteers, not by a funded research department. The milestones below reproduce the institution's published "Our Journey" timeline. [27]
| Year | Milestone (as published by Awqaf NZ) |
|---|---|
| 2011 | Awqaf New Zealand is founded (February), driven by the vision of a sustainable Qurbani industry. Within six months, it signs its first Protocol of Cooperation with an established national awqaf institution. |
| 2012 | Establishes that owning dedicated livestock farms is essential to building a sustainable Qurbani supply chain. |
| 2013 | Designs the Smart Waqf Fund model and wins the Islamic Economy Award (Awqaf Category); congratulated by the Secretary-General of a national awqaf authority. |
| 2014 | Identifies sheep-dairy farming as the ideal model for the Qurbani industry, each ewe generating value across a five-year lifecycle before Qurbani. |
| 2015 | Meets the Securities Commission Malaysia to explore issuing an Awqaf Sukuk through the Smart Waqf model. |
| 2016 | Commissions ISRA to verify the Smart Waqf model, and begins a series of international Awqaf industry workshops (Kuala Lumpur, Dubai, Madinah). |
| 2017 | Completes five international Awqaf industry workshops (adding Edmonton and Kuala Lumpur) and opens international discussions on the fiqh distinction between Temporary Cash Waqf and Qard Hasan (Indonesia). |
| 2018 | Identifies the critical fiqh distinction between Temporary Cash Waqf and Qard Hasan for drafting Awqaf finance contracts; takes part in the IMF and World Bank meetings in Bali, where the Waqf Core Principles (WCP) are launched (the global waqf-governance framework in whose working group Awqaf NZ served). |
| 2019 | Presents the first draft of the Fiqh Differences study at the 15th International Forum of the Maliki School (authored by Dr Abdul Qadar Qadawi, Algeria). |
| 2020 | Signs a Memorandum of Understanding with a national awqaf authority. |
| 2021 | Renews that Memorandum of Understanding, advancing the fiqh of Temporary Cash Waqf and the first legal drafts of the associated contracts. |
| 2022 | Develops a faith-based marketing strategy for cash waqf across diverse global communities. |
| 2023 | Drafts the first Temporary Cash Waqf contracts, grounded in the Fiqh Differences study. |
| 2024 | Presents Awqaf NZ's Smart Waqf experience at the 16th International Forum of the Maliki School, Algeria. |
| 2025 | Visits a national awqaf authority for Shariah and technical support; joins an international seminar on digital endowments; and meets a national waqf body regarding the Qurbani sheep-dairy programme. |
| 2026 | The Year of AI: harnessing artificial intelligence to overcome the limits of a volunteer team, launching a new multilingual website, and preparing for a first practical sheep-dairy pilot. |
The 2018 entry is supported externally: the published Waqf Core Principles, hosted by BWI, list Dr Husain Benyounis of Awqaf New Zealand as a member of the International Working Group (the same list includes Dr Hisham Dafterdar of Awqaf Australia, whose essay prompted this report), and Awqaf NZ's records show participation in the working group's meetings and the Bali launch in October 2018. [33-34]
This history changes how 2026 should be understood. Calling 2026 the Year of AI does not mean that AI discovered the legal questions, wrote the institutional history or carried out the human consultations. AI helped volunteers find, organise, translate, criticise and communicate work accumulated over fifteen years. It did not settle the Temporary Cash Waqf and Qard Hasan question. That question is now the first topic of the 12th Waqf Fiqh Issues Forum of the Kuwait Awqaf Public Foundation, whose ruling Awqaf NZ awaits, and no capital will be sought until a ruling is given. The 2025 entry's "international seminar on digital endowments" is the Kuala Lumpur seminar described in Section 4.6. [39-40]
By November 2026 Awqaf NZ's public record included a website in six languages (English, Arabic, French, German, Indonesian and Turkish); a series of openly licensed research publications, several deposited on Zenodo with DOIs and gathered in an Awqaf New Zealand community there, including the Qurbani Terms and Conditions framework, the Agriculture 4.0 working paper, a vocational education and training paper, a sheep-dairy feasibility study and a discussion paper on Hijri-aligned sheep dairying; the first indexing of its 2011 to 2018 international archive (1,211 files); and multilingual Animal Family educational media. These are counts of outputs, not measures of impact. [25-27,35-38]
| Workstream | 2026 contribution of AI | Evidence boundary |
|---|---|---|
| Research | Faster multilingual searches, comparison, outlines and drafting | Publication is not peer review or expert approval |
| Archive recovery | Indexing and retrieval of historical institutional files | Retrieval does not prove every summary |
| Agriculture 4.0 and sheep dairy | Literature review, feasibility and discussion papers | Proposals, not an operating farm or measured impact |
| Qurbani terms review | Comparative examples and a proposed rights-based framework | Exploratory sample, not a universal audit |
| Website and correspondence | Six-language site, letters and institutional correspondence | Human approval of every published statement |
| Visual education | Multilingual cartoons and infographics | Educational scenes are not records of services |
| International cooperation | Sharing the approach with sister Awqaf bodies in Canada, France and Finland | Knowledge sharing, not completed replication |
This status discipline is a practical control against the tendency of powerful drafting tools to turn "we plan" into "we deliver", or "we are discussing" into "we have partnered".
The tools were used for distinct purposes, each under human direction:
The strongest method was not a competition among products but a division of labour with a shared human review point. One tool might find a source, another might identify a missing consideration; a volunteer then inspected the source, labelled uncertainties and decided whether the text should stay. Outputs were versioned and marked by approval status. Product names are given for transparency, not as endorsement or sponsorship.
For a volunteer institution, data governance is concrete. Files include trustee correspondence, drafts, contracts, archives, partner exchanges and, if a service is ever implemented, information about offerors and recipients. Scattered personal accounts make permissions and continuity hard to manage. Awqaf NZ has moved its administration into a Google Workspace for Nonprofits environment, while using other AI providers for appropriately scoped public research, editorial work and design. The environment is a control opportunity, not evidence of a completed privacy or compliance audit.
What the provider documents establish. Google's documentation for the no-cost nonprofit edition lists Shared Drives, central administration and the Gemini app with enterprise-grade data protections, under which chats and uploaded files are not reviewed by humans or used to train models. Google's Workspace terms state that customer data are not used to train covered generative AI models without permission, and its Privacy Hub explains that Workspace Gemini services respect existing access permissions. These are statements under defined terms and editions, not assurances that every Google service, connected app or organisation-specific duty is covered. [28-31]
The edition matters. Google's comparison lists Google Vault and Data Loss Prevention (DLP) only for paid editions, and in-app Gemini features (such as the Gemini side panel in Gmail, Docs and Drive) are reserved for paid editions rather than the no-cost nonprofit edition. Google also distinguishes the Gemini app, Gemini in Workspace and Gemini Notebook (formerly NotebookLM), each with its own data-handling arrangements: for example, direct uploads and notebooks in Gemini Notebook persist until deleted, while sources imported from Google Drive respect existing Drive view permissions, sync with the originals and are cached by Gemini Notebook for performance; the organisation's data-region settings do not apply to that cached data. Gemini Notebook is a core Workspace service in the nonprofit edition, whereas optional connected Google services (such as Maps or YouTube) are governed by separate Additional Product Terms. [32,42] This report therefore does not describe Awqaf NZ as using retention, audit or DLP features it does not have, and does not claim enterprise parity at no cost. [28,32]
| Control question | Practical approach | Verification needed |
|---|---|---|
| Where do master records live? | Organisationally managed Drive and Shared Drives, with clear ownership and versioning | Actual folders, external sharing, backups and access lists |
| Who can use organisational AI? | Designated trustees and volunteers, with two-step verification and periodic access review | Which features the edition actually includes |
| Which AI receives which information? | Public or de-identified research with approved external tools; confidential material only after separate approval | Consumer and connected services may have different terms |
| Can Gemini see restricted records? | It follows Workspace access permissions within supported workflows | A wrongly shared file is a human permissions failure |
| What retention and audit exist? | Product settings plus a simple manual control log | No Vault or enterprise DLP in the free edition |
| Who certifies compliance? | Awqaf NZ's authorised people review terms, privacy notices, consents and processes | Provider claims do not certify Awqaf NZ's compliance |
The safeguard is organisational, not brand-dependent: keep records in a managed repository, separate internal data from exploratory prompts, grant least-privilege access, check each enabled service and agreement, and keep a human approval trail. Included licences may reduce software fees, but volunteer time, training, review and security still cost something. Google is a technology supplier, not a programme partner, sponsor or Shariah endorser, and sister organisations must check their own eligibility, edition, terms and national law. [28-32]
Awqaf NZ's volunteers wanted to understand how Qurbani providers worldwide describe contractual authority, animal requirements, fulfilment, recipients, worker protections, food safety and the use of hides, wool and other resources. The initial register coded 29 publicly accessible Terms and Conditions documents from several Western countries, with Arabic and other-language expansion proposed. The register is exploratory and not a representative sample until its inclusion criteria and coding are independently checked. [25]
The challenge was not only translation. Similar-looking terms in different languages can encode different legal relationships or religious assumptions. A one-language search would miss material sources, while a manual review of comparable breadth would have exceeded a volunteer team's time and language capacity.
| Stage | AI-assisted task | Human requirement |
|---|---|---|
| Discover | Locate public terms across languages and jurisdictions | Record URL, capture date, version and provider |
| Extract | Identify clauses and service descriptions | Check passages against the original, not a summary |
| Classify | Group issues by rights-holder, stage and obligation | Review ambiguous classifications and translations |
| Compare | Flag missing or inconsistent provisions within the sample | Do not infer that omission proves illegality |
| Interpret | Draft a hypothesis about contract transparency | Seek local legal and Shariah interpretation |
| Propose | Develop a rights-based framework and explanatory material | Mark as a proposal and publish limitations |
| Review | Invite scholarly, legal and operational critique | Record decisions, corrections and open questions |
The comparison suggested that many sampled documents concentrated on price, animals or shares and fulfilment dates, with less visibility of wider rights and responsibilities. This supported a research proposition: Qurbani services could benefit from a clearer international framework. It is a proposal, not proof that the global Qurbani industry lacks regulation; consumer, charity, veterinary, employment, public-health and food law may apply even where a provider's terms are incomplete. [25]
The proposed framework identifies nine connected interests: the offeror (al-mudhahhi); service organisations; animals; recipients; workers; food safety; public health; the environment; and resource stewardship, including hides, wool and other materials. It separates (1) common service terms, (2) a Shariah governance annex approved for the relevant arrangement, and (3) jurisdiction-specific legal provisions, so that no single generic contract claims universal sufficiency. [25]
AI did not decide who owns the hide, issue a fatwa, certify a slaughter method, identify recipients or authorise a service. Those remain questions for the appropriate people and authorities.
The same method could be applied, carefully, to Awqaf governance policies, service contracts, grant conditions, educational materials or translated disclosures. The transferable contribution is a repeatable way to surface questions and documentary gaps, followed by verification, not the specific conclusion reached in any one domain.
The following principles are drawn from Awqaf NZ's experience and offered as a working approach, not an adopted code or certified policy.
9.1 Faith, purpose and amanah first. Ask what entrusted purpose will benefit, who may be harmed and which obligations cannot be displaced. No technology should be adopted merely because it is available.
9.2 Human and scholarly authority remain identifiable. Trustees, managers, specialists and qualified scholars approve matters within their competence. A machine-generated fiqh answer, legal assurance or operational commitment cannot replace the institution's duty of care. [19-21,24]
9.3 Protect data before asking the model. Classify files as public, internal, confidential or sensitive; use de-identified material where possible; keep identifiable offeror or recipient information out of unsuitable tools; and check terms, permissions and privacy law before using a new service. [10-11]
9.4 Keep evidence and intention apart. History, verified facts, assumptions, proposals, approvals, live operations and audited results need explicit labels. A design diagram does not prove a system is operating; an MoU discussion does not prove a contract is signed.
9.5 Inspect primary sources. Citations should be retrievable, translations read against originals, numbers reconciled, and legal or Shariah interpretations checked by the right professional. Agreeing models are not independent verification.
9.6 Disclose AI assistance honestly. State what AI did, what was checked, who holds editorial responsibility and what remains under review. Name products factually without implying endorsement.
9.7 Respect volunteers' work and limits. Credit the people who built, interpreted and approved the research, and budget for verification: reading, legal checks, language review, training and corrections take time even when software is free.
9.8 Preserve the human relationship. Offerors, recipients and partners must be able to reach a responsible person. An unsupervised chatbot or an attractive synthetic image is no substitute.
9.9 Learn from corrections. Record errors and how they were found. Counting errors caught is as useful as counting pages drafted.
9.10 Replicate the method, not the legal status. Sister Awqaf bodies in Canada, France and Finland may adapt research methods, tool registers, review logs and education practices, but they do not inherit Awqaf NZ's mandates, permissions, contracts or regulatory status.
Disclosure helps only when readers can tell what a tool did, what a person checked and what remains provisional. The following model notices are offered for adaptation and translation; any statement about verification must match the actual review record. [6,20]
Awqaf New Zealand uses selected artificial intelligence tools to help volunteers with public-source research, translation, document organisation and educational design. People remain responsible for editorial approval and for all trustee, Shariah, legal, financial and operational decisions. Some materials describe research proposals rather than operating programmes. Use of a technology provider does not imply its sponsorship or endorsement.
AI assistance and verification: this document was prepared with assistance from [tools actually used] for [research, comparison, editing, translation or design]. Human reviewers checked sources and specialist claims to the extent stated. Matters not yet verified, approved or tested are identified as such. Awqaf New Zealand retains editorial responsibility. This document is not a fatwa, legal opinion or approved service unless stated.
Educational illustration notice: the Animal Family and related scenes are fictional, sometimes AI-assisted illustrations. They explain stewardship concepts and do not depict real beneficiaries, animals allocated to a service, completed Qurbani distributions or verified operations. Service information and evidence of completion are provided separately through authorised records.
Model D: Google Workspace reference
Awqaf New Zealand uses Google Workspace for Nonprofits as an internal administrative and research environment, subject to the products and agreements applicable to its account. Google does not sponsor, endorse or operate Awqaf NZ programmes. Personal information is handled under Awqaf NZ's own privacy and access policies, and AI uses are assessed by the sensitivity of the data.
Translations should preserve the distinctions between draft, proposed, approved, verified and operating. A notice does not replace checking the safety of the underlying workflow; under Gate 5 below, the release checklist records which notice was used and who approved it.
The model below is offered for testing. It gives a board with few or no staff a practical way to make decisions proportionate to risk, combining experience-based controls with the emphasis on human oversight, privacy, transparency and documentation in regulator guidance and UNESCO's recommendation. [8-12]
| Gate | Question to answer | Minimum record | Responsible |
|---|---|---|---|
| 1. Purpose and risk | Why use AI, and what harm could follow? | Task brief, expected benefit, risk tier | Project lead with trustee oversight |
| 2. Tool and data authority | Is this product and edition approved for these data and terms? | Tool register, edition, data class, access and retention rules | Account administrator or privacy lead |
| 3. Source verification | Which factual, legal or Shariah claims are traceable? | Source register, exceptions, reviewer notes | Research lead or specialist |
| 4. Human decision | Who has authority to approve this conclusion or action? | Signed review or meeting record | Qualified decision-maker |
| 5. Controlled release | Are status, version, AI roles and permissions disclosed? | Release checklist, approved notice (9.11), consents | Editor or board delegate |
| 6. Measure and correct | Did it save time, improve quality or cause harm? | Correction log, outcomes, cost and feedback | Team lead and periodic trustee review |
| Tier | Illustrative tasks | Minimum safeguards | Default |
|---|---|---|---|
| A, low | Indexing public documents, outlines, early ideas | Approved tool, source link, human screening | Pilot with normal controls |
| B, moderate | Translation, website copy, cartoons, explanations | Language review, disclosure, factual checks | Release after editorial approval |
| C, high | Legal or Shariah research, personal records, contract terms, beneficiary analysis | Specialist supervision, privacy assessment, documented evidence, restricted access | No autonomous operational decision |
| D, reserved to humans | Binding fatwa, trustee approval, final beneficiary entitlement, signing contracts, banking mandates | Authorised people and procedures; AI may help prepare | No delegation of final authority |
These tiers are an Awqaf NZ proposal, not a legal classification. National law may require stronger controls, and some uses may be prohibited regardless of their apparent usefulness.
A one-page entry can record: task and owner; intended beneficiaries; data class; AI tool and account type; source links; material errors found; specialist referrals; consent status; final approver; version and release date; and corrective action. It should be short enough that volunteers use it every time, and can be kept in the institution's managed environment, with the edition and permissions checked as in Section 7.4.
A pilot report could measure research questions answered, verifiable citations per report, errors caught in review, hours of human review, translation corrections, files made retrievable, reader comprehension and total cost, and then ask whether the work changed a decision or improved a service. Counting pages and videos is easy; showing better governance and social benefit is harder, and more important.
Sister Awqaf organisations are invited to start with a small, reversible pilot before any sensitive or transaction-bearing automation. Each institution keeps its own independence and approvals.
| Period | Activity | Output and decision point |
|---|---|---|
| Days 1 to 15 | Choose a public-information task; nominate two reviewers; write a one-page use policy | Approved task, risk tier and named reviewers |
| Days 16 to 30 | Build an approved tool list; set data boundaries; train volunteers on sources and errors | Tool and privacy register; examples of unacceptable prompts |
| Days 31 to 60 | Run a small multilingual research or archive pilot against a manual baseline | Search log, verified sources, time spent and errors found |
| Days 61 to 75 | Review with trustees and relevant specialists | Decision to adopt, change or stop; open issues recorded |
| Days 76 to 90 | Publish only an approved, disclosed example; share templates with peers | Versioned output, reader feedback and next-cycle plan |
The most technically confident volunteer should not automatically become the authority on Shariah, contracts, veterinary questions or beneficiary dignity; equally, a trustee need not learn model engineering to require proof of source checking and privacy compliance. Proportional governance means clear roles, not expensive bureaucracy.
Five questions deserve collaborative study. First, does multilingual AI research find materially more relevant sources than conventional methods once verification costs are counted? Second, what citation, accuracy and translation-error rates arise in fiqh-adjacent research? Third, how do offerors and beneficiaries respond to transparent fictional cartoons compared with human-led information? Fourth, what minimum privacy and data controls can volunteer teams sustain? Fifth, what evidence would show that transferring Awqaf NZ's method improved outcomes elsewhere, rather than simply increasing document output?
Future studies should pre-register search terms and jurisdictions, keep dated source registers, define coding categories, record human adjudication of ambiguous cases, and separate labour saved from new review labour. Experiments should be designed with scholars, lawyers, operations specialists and affected communities, moving the sector from enthusiasm and frameworks towards evidence about effectiveness and harm.
The limitations are deliberate. The international review is purposive, not exhaustive; survey figures from different countries are not a pooled measure of the Awqaf sector; public AI tools and strategies were not functionally tested; the account of Awqaf NZ's work rests on its own records and publications; the Google Workspace discussion is not an audit of account settings or data flows; and the Qurbani sample cannot establish the absence of law or the prevalence of breaches. The report has not undergone formal Shariah, legal or peer review.
AI can help volunteer-led Awqaf institutions cross real barriers of language, research time, scattered archives and scarce specialist capacity. Awqaf NZ's 2026 experience shows how several tools can work together under human direction to turn dispersed information into documents that are easier to examine, challenge and communicate. The Qurbani Terms and Conditions example matters not because AI announced a universal truth, but because it helped the team ask better questions about rights, stewardship and responsibility.
The international evidence supplies the counterweight. Adoption is not governance; productivity is not mission impact; a public tool is not independently verified; and an AI output never substitutes for the people entrusted with religious and fiduciary authority. The aim is neither enthusiasm without controls nor prohibition without examination, but disciplined, proportionate and transparent use: greater capability paired with greater accountability. Awqaf NZ offers this proposition for public discussion and practical testing.
AI assistance. This report was prepared largely by AI tools, and Awqaf NZ states this openly. The author set the questions, supplied the institutional record, chose what to include and approved the text; the AI tools searched, translated, drafted, cross-checked and typeset. It developed from a reply to Dr Dafterdar's essay drafted with Claude (Anthropic). Drafts 1 and 2 were researched and written with ChatGPT (OpenAI), with review observations from Gemini (Google). Draft 3 and this Version 1.0 were revised, source-checked and typeset with Claude through Cowork, with Gemini reviewing the report's statements about Google Workspace for Nonprofits against Google's published terms. That was an AI review, not certification by Google or a legal audit. For the comparative sections, Gemini and ChatGPT each ran the same structured searches in several languages, and Claude checked their claims against the original sources, rejecting or flagging those that could not be confirmed: the institutional timeline was aligned with Awqaf NZ's published record, survey figures and institutional descriptions were checked against their published originals as at 21 November 2026, and unverifiable statements were corrected or removed. Canva contributed to the educational media described in Section 7, not to this manuscript. The tools are named for transparency, not as sponsors, co-authors, endorsers or Shariah authorities. The author and Awqaf New Zealand remain responsible for the content.
Disclaimer. This research report is published for public and scholarly discussion. It is not a fatwa, legal opinion, peer-reviewed article, adopted AI governance standard, financial solicitation or Qurbani service offer, and its governance cycle is a proposal for testing. Statements about future services and cooperation are not evidence of operational implementation or third-party consent. No technology provider, regulator, university, Awqaf body or person named here endorses this report. Dr Hisham Dafterdar's essay is cited as a public contribution to the discussion; he is not a co-author and has not approved this report.
Sources accessed 21 November 2026 unless stated otherwise.
[1] State of Nonprofit AI Adoption and Governance
NTEN and The Bridgespan Group. 10 September 2026.
Links: https://www.nten.org/publications/state-of-nonprofit-ai
[2] The State of Artificial Intelligence Adoption in Canadian Nonprofits
Imagine Canada and partners. 20 January 2026.
Links: https://imaginecanada.ca/en/research/ai-report
[3] L'usage de l'IA au sein des OSBL faisant appel aux générosités, volet 1 : pratiques de l'organisation et des individus
France générosités. June 2026 (published 9 June 2026).
Links: https://www.francegenerosites.org/wpcb/download/103/documents-grand-public/47931/usage-de-lia-au-sein-des-osbl-volet-1-etude-france-generosites-2026.pdf
[4] Guide : Encadrer l'usage de l'IA au sein des associations et fondations
France générosités and Don en Confiance. 17 March 2026.
Links: https://www.donenconfiance.org/offres/gestion/events_759_57153_non-1813/guide-encadrer-l-usage-de-l-ia-au-sein-des-associations-et-fondations.html
[5] Künstliche Intelligenz in der Stiftungsarbeit (Stiftungsfokus Nr. 22)
Bundesverband Deutscher Stiftungen. 2026. Based on the Stiftungspanel annual survey 2026 of 450 foundations.
Links: https://www.stiftungen.org/fileadmin/stiftungen_org/Verband/Was_wir_tun/Publikationen/Stiftungsfokus-22-KI.pdf
[6] Bots for good? What people in the UK think about charities using AI
YouGov, J. Fernandes. 19 November 2026. 2,097 GB adults; fieldwork 19 to 20 August 2026.
Links: https://yougov.com/en-gb/articles/56013-bots-for-good-what-people-in-the-uk-think-about-charities-using-ai
[7] AI With Purpose: How Foundations and Nonprofits Are Thinking About and Using Artificial Intelligence
Center for Effective Philanthropy. 2025.
Links: https://cep.org/report-backpacks/ai-with-purpose-how-foundations-and-nonprofits-are-thinking-about-and-using-artificial-intelligence/
[8] Charities and artificial intelligence
Australian Charities and Not-for-profits Commission. Official guidance.
Links: https://www.acnc.gov.au/tools/guides/charities-and-artificial-intelligence
[9] Charity Sector Risk Assessment 2026
Charity Commission for England and Wales. 18 August 2026.
Links: https://www.gov.uk/government/publications/charity-sector-risk-assessment-2026/charity-sector-risk-assessment-2026
[10] Generative Artificial Intelligence
Office of the Privacy Commissioner, New Zealand. Official guidance.
Links: https://www.privacy.org.nz/resources-and-learning/a-z-topics/ai/generative-artificial-intelligence/
[11] What is in a good AI Policy? and AI policy templates
CommunityNet Aotearoa and NZ Navigator Trust. April and May 2026.
Links: https://community.net.nz/resources/nz-navigator-trust/ai-policy · https://community.net.nz/resources/nz-navigator-trust/ai-policy-templates
[12] Recommendation on the Ethics of Artificial Intelligence
UNESCO. Adopted 23 November 2021.
Links: https://www.unesco.org/en/legal-affairs/recommendation-ethics-artificial-intelligence
[13] Literasi Wakaf AI Assistant
Badan Wakaf Indonesia. Official public service.
Links: https://www.bwi.go.id/literasi-wakaf-ai-assistant/
[14] تمكين القطاع الوقفي (Enabling the Awqaf sector)
General Authority for Awqaf, Saudi Arabia. Official sector guidance.
Links: https://web.awqaf.gov.sa/ar/%D8%AA%D9%85%D9%83%D9%8A%D9%86-%D8%A7%D9%84%D9%82%D8%B7%D8%A7%D8%B9-%D8%A7%D9%84%D9%88%D9%82%D9%81%D9%8A
[15] MAIWP Launches 2026-2030 Strategic Plan With AI Integration
Bernama. 14 May 2026.
Links: https://www.bernama.com/en/general/news.php?id=2557067
[16] Waqf, Dakwah Bodies Must Adapt To AI, Digitalisation Without Losing Values: PM Anwar
Bernama. 27 July 2026.
Links: https://www.bernama.com/en/news.php?id=2586626
[17] 12th International Conference on Mosque, Zakat and Waqf (IMAF 2026): Digital Transformation and Artificial Intelligence in the Management of Mosques, Zakat and Waqf
Universiti Islam Selangor. 15 September 2026, Bandar Baru Bangi.
Links: https://conference.uis.edu.my/i-maf/
[18] Vakıf Faaliyetlerinde Yapay Zekâ Kullanımı: Hukuki Düzenleme ve Uygulama Örnekleri
Burcu Şenlikoğlu. Tokat Gaziosmanpaşa Üniversitesi Hukuk Fakültesi Dergisi (TOGÜHFD), 3(1), 2025.
Links: https://dergipark.org.tr/tr/download/article-file/4530718
[19] حكم استخدام تطبيقات الذكاء الاصطناعي (AI) في الاستفتاء (Ruling on using AI applications to seek fatwas)
Dar al-Ifta al-Misriyyah. 22 September 2025.
Links: https://www.dar-alifta.org/ar/Fatwa/Details/22158/
[20] حكم استخدام الذكاء الاصطناعي (AI) في كتابة المحتوى العلمي ثم نسبته إلى النفس (Ruling on using AI to write scholarly content and attributing it to oneself)
Dar al-Ifta al-Misriyyah. 2025.
Links: https://www.dar-alifta.org/ar/Fatwa/Details/22203/
[21] حكم الاعتماد على التقارير الطبية الصادرة عن الذكاء الاصطناعي (Ruling on relying on AI-generated medical reports)
Dar al-Ifta al-Misriyyah. Fatwa 8897, 3 March 2026.
Links: https://www.dar-alifta.org/ar/Fatwa/Details/22614/
[22] AI Is Coming to Waqf and Zakat: Are Islamic Social Finance Institutions Ready? A Governance-First Conceptual Framework for Responsible Artificial Intelligence Adoption (GF-AIRISF)
Ibrahim Abdul Mugis, Awqaf Africa. 2026. Conceptual framework.
Links: https://awqafafrica.org/ai-readiness-waqf-zakat-gf-airisf/
[23] When May AI Decide? Shariah, Maqasid and the Limits of Artificial Intelligence Delegation in Waqf, Zakat and Islamic Social Finance
Ibrahim Abdul Mugis, Awqaf Africa. Global WaqfTech Research Paper No. 02, 2026. Conceptual paper.
Links: https://awqafafrica.org/when-may-ai-decide-shariah-maqasid-ai-waqf-zakat/
[24] Awqaf and AI Governance: Balancing Faith and Technology
Hisham Dafterdar. Essay posted publicly in the Global Awqaf Forum WhatsApp group, November 2026. Dr Dafterdar is a Trustee of Awqaf New Zealand and Chairman of Awkaf Australia Ltd.
[25] Towards an International Framework for the Terms and Conditions of Qurbani Services
Husain Benyounis. Awqaf New Zealand, Version 1.0, 8 November 2026.
Links: https://doi.org/10.5281/zenodo.24042729
[26] Securing Qurbani Livestock and Building Waqf Capability: An objectives-led assessment of Agriculture 4.0 for a proposed sheep-dairy pilot
Husain Benyounis. Awqaf New Zealand Working Paper, Version 1.2, 20 November 2026.
Links: https://doi.org/10.5281/zenodo.24394096
[27] About Awqaf New Zealand: Our Journey 2011 to 2026
Awqaf New Zealand. Official website.
Links: https://www.awqafnz.org/en/about.html
[28] Compare Google Workspace features for nonprofits
Google. Official feature comparison.
Links: https://knowledge.workspace.google.com/admin/billing/compare-google-workspace-features-for-nonprofits
[29] Google Workspace for Nonprofits edition
Google for Nonprofits. Official help.
Links: https://support.google.com/nonprofits/answer/2858465
[30] Generative AI in Google Workspace Privacy Hub
Google Workspace. Official privacy documentation.
Links: https://knowledge.workspace.google.com/admin/generative-ai/generative-ai-in-google-workspace-privacy-hub
[31] Google Workspace Service Specific Terms
Google Workspace. Including the generative AI training restriction.
Links: https://workspace.google.com/terms/service-terms/
[32] Google Workspace Additional Product Terms
Google Workspace. Terms for optional additional Google services, such as Maps or YouTube.
Links: https://workspace.google.com/terms/additional_services/
[33] Core Principles for Effective Waqf Operation and Supervision (Waqf Core Principles)
Bank Indonesia, Badan Wakaf Indonesia and IRTI-IsDB, International Working Group on Waqf Core Principles. October 2018. The members list includes Dr Husain Benyounis, Awqaf New Zealand.
Links: https://www.bwi.go.id/wp-content/uploads/2019/10/102018-Core-Pinciples-for-Effective-Waqf-Operation-and-Supervision.pdf
[34] Indonesia Luncurkan Waqf Core Principles dan Waqf Linked Sukuk di Forum IMF dan Bank Dunia
Badan Wakaf Indonesia. 14 October 2018.
Links: https://www.bwi.go.id/1641/2018/10/14/indonesia-luncurkan-waqf-core-principles-dan-waql-linked-sukuk-di-forum-imf-bank-dunia/
[35] Vocational Education and Training as an Engine of Awqaf Industry Development
Husain Benyounis. Awqaf New Zealand Working Paper, Version 1.1, 16 November 2026.
Links: https://doi.org/10.5281/zenodo.24253676
[36] Feasibility Study: Qurbani Sheep-Dairy Pilot, Franklin / North Waikato, New Zealand (Draft v4 for review)
Awqaf New Zealand. October 2026.
Links: https://doi.org/10.5281/zenodo.23279953
[37] Hijri-Aligned Qurbani Sheep-Dairy Farming: The Economy of the Qurbani Ritual and an Open Invitation to Research, Test and Improve a Proposed Livestock Production System
Husain Benyounis. Awqaf New Zealand Discussion Paper No. 1, Version 1.0, 30 September 2026.
Links: https://doi.org/10.5281/zenodo.23044285
[38] Awqaf New Zealand community
Zenodo. Collection of Awqaf New Zealand research deposits.
Links: https://zenodo.org/communities/awqafnz
[39] توصيات ندوة الأوقاف الرقمية: مشروعيتها، تطبيقاتها، وأثرها في التنمية المستدامة (Recommendations of the Seminar on Digital Endowments: Their Legitimacy, Applications and Role in Sustainable Development)
Kuwait Awqaf Public Foundation, International Islamic Charity Organization and International Forum of Development Organizations, with the International Islamic University Malaysia. Kuala Lumpur, 18 to 19 February 2025. Adopted recommendations and seminar programme. No public online copy located; copy supplied to the author and available on request, subject to the publisher's permission.
[40] الموضوعات وشروط تقديم الأبحاث لمنتدى قضايا الوقف الفقهية الثاني عشر: قضايا مستجدة وتأصيل شرعي (Topics and submission conditions for the 12th Waqf Fiqh Issues Forum)
Kuwait Awqaf Public Foundation, Department of Studies, Information and Documentation. Introductory booklet, 2025. No public online copy located; copy supplied to the author and available on request, subject to the publisher's permission.
[41] أوراق عمل ندوة الأوقاف الرقمية (Working papers of the Seminar on Digital Endowments)
Kuwait Awqaf Public Foundation and partners, Kuala Lumpur, February 2025. Cited: Sara Mutla Al-Qahtani, "Digital Waqf between the problem of concept and Shariah characterisation: an epistemological Shariah approach", working paper 1, pp. 4-20; Omran Salem, "Digital endowments: technical features, challenges and prospects for development in the digital age", working paper 8, pp. 124-140. The copy consulted has 200 pages, while its contents list papers beginning on pp. 208-209, so it may not be the complete proceedings. No public online copy located; copy supplied to the author and available on request, subject to the publisher's permission.
[42] Turn Gemini Notebook on or off for users
Google Workspace Admin Help. Including Drive caching, data regions and core-service editions.
Links: https://knowledge.workspace.google.com/admin/generative-ai/gemini-notebook/turn-gemini-notebook-on-or-off-for-users
[43] Use apps connected to Gemini with a work or school Google Account
Google Help. Gemini app connection to Gmail and Drive.
Links: https://support.google.com/gemini/answer/14959807
[44] Management of Wakaf Assets
Ministry of Culture, Community and Youth, Singapore. Response to parliamentary question, 26 February 2026.
Links: https://www.mccy.gov.sg/about-us/news-and-resources/management-of-wakaf-assets/
[45] Strengthening Wakaf: Wakaf Revitalisation Scheme and asset migration
Majlis Ugama Islam Singapura (MUIS). Official page.
Links: https://www.muis.gov.sg/community/wakaf-administration/strengthening-wakaf/
[46] Speech by Chief Executive of MUIS, Mr Kadir Maideen, at Wakaf Disbursement 2026
Majlis Ugama Islam Singapura (MUIS). 24 July 2026.
Links: https://www.muis.gov.sg/resources/speeches/speech-by-chief-executive-of-muis--mr-kadir-maideen-at-wakaf-disbursement-2026/
[47] Use of Generative AI by Muslims in Singapore
Ministry of Culture, Community and Youth, Singapore. Response to parliamentary question, 26 February 2026.
Links: https://www.mccy.gov.sg/about-us/news-and-resources/use-of-generative-ai-by-muslims-in-singapore/
[48] Acting Minister Zaqy sets next direction for MUIS
Majlis Ugama Islam Singapura (MUIS). Media release, 4 August 2026.
Links: https://www.muis.gov.sg/resources/media-releases/acting-minister-zaqy-sets-next-direction-for-muis/
[49] First series of Fatwa Lab
Majlis Ugama Islam Singapura (MUIS), Office of the Mufti. Including the research topic on optimising fatwa research with AI.
Links: https://www.muis.gov.sg/resources/khutbah-and-religious-advice/fatwa-lab/first-series-of-fatwa-lab/
[50] Account of an AI4EDS train-the-trainer workshop with Warees Investments
LinkedIn post by an AI Singapore trainer, 2026. Self-reported; not independently verified for this report.
Links: https://sg.linkedin.com/in/sri-bbb-21142b61
[51] Sharing experiences and ideas: Dompet Dhuafa and Warees Investment Singapore strengthen waqf management and digital transformation
Dompet Dhuafa. 27 August 2026.
Links: https://www.dompetdhuafa.org/en/sharing-experiences-and-ideas-dompet-dhuafa-and-warees-investment-singapore-strengthen-waqf-management-and-digital-transformation/
[52] Awqaf General Directorate, Wahed MENA seal MoU to develop AI platform for Sharia-compliant equity analysis
Qatar News Agency (QNA). 10 June 2026.
Links: https://qna.org.qa/en/News-Area/News/2026-6/10/awqaf-general-directorate-wahed-mena-seal-mou-to-develop-ai-platform-for-sharia-compliant-equity-analysis
[53] Ministry of Awqaf announces strategic partnership with Google Cloud
The Peninsula (QNA). 13 December 2025.
Links: https://thepeninsulaqatar.com/article/13/12/2025/ministry-of-awqaf-announces-strategic-partnership-with-google-cloud
[54] Ministry of Awqaf launches "Live Fatwas" service on Islamweb
Qatar News Agency (QNA). 8 September 2025. See also: Awqaf Ministry launches AI-powered FatwaTok app, The Peninsula, 2 March 2026.
Links: https://qna.org.qa/en/News-Area/News/2025-9/8/ministry-of-awqaf-launches-live-fatwas-service-on-islamweb · https://thepeninsulaqatar.com/article/02/03/2026/awqaf-ministry-launches-ai-powered-fatwatok-app
[55] Katara Tech Forum highlights AI's role in Islamic Sharia applications
Qatar Tribune. 8 July 2026.
Links: https://www.qatar-tribune.com/article/243434/latest-news/katara-tech-forum-highlights-ais-role-in-islamic-sharia-applications
[56] Awqaf employees to be trained in artificial intelligence
Khaleej Times. 29 October 2025.
Links: https://www.khaleejtimes.com/uae/awqaf-employees-trained-artificial-intelligence
[57] Awqaf Abu Dhabi unveils its first photoreal AI agent at "Ai Everything Abu Dhabi"
Emirates News Agency (WAM). October 2026.
Links: https://www.wam.ae/en/article/c2lun1e-awqaf-abu-dhabi-unveils-its-first-photoreal-agent
[58] UC Investments 2024-25 Annual Report
Office of the Chief Investment Officer of the Regents, University of California.
Links: https://www.ucop.edu/investment-office/2024-25-annual-report.pdf
[59] Rice University Endowment Report, Fiscal Year 2025
Rice Management Company. Published March 2026. Section "Artificial Intelligence: A Reality in Portfolio and Operations".
Links: https://giving.rice.edu/sites/g/files/bxs4641/files/2026-03/FY25%20Report%20for%20Website.pdf
[60] Capital Allocators podcast: AI in the investment office
Global Endowment Management (GEM). 16 September 2026.
Links: https://www.geminvestments.com/insights/mbb-capital-allocators-podcast/
[61] Cambridge University Endowment Fund Annual Review 2025
University of Cambridge Investment Management Limited.
Links: https://www.ucim.co.uk/annual-report-2025/overview
[62] The AI opportunity for investment analysis: Oxford University Endowment Management
Thorogood. Case study, about December 2024.
Links: https://www.thorogood.com/perspectives/case-studies/tgcas202iqr-the-ai-opportunity-for-investment-analysis/
[63] Board of Directors meeting book, 18 June 2026
University of Texas/Texas A&M Investment Management Company (UTIMCO). Minutes of 5 March 2026 record an overview of AI upskilling.
Links: https://www.utimco.org/media/53kfyh0x/book-board-of-directors-june-18-2026_for-web.pdf
[64] Legal Operations Analyst (job advertisement)
Cascade Asset Management Company, manager of the Gates Foundation Trust's investments. November 2026.
Links: https://www.legal.io/jobs/5886044/Full-time/Legal-Operations-Analyst/Kirkland/Washington
[65] Master's student capstone spotlight: AI-enabled information extraction for investment management
Harvard John A. Paulson School of Engineering and Applied Sciences. 10 June 2024.
Links: https://seas.harvard.edu/news/masters-student-capstone-spotlight-ai-enabled-information-extraction-investment-management
[66] KI und Gesellschaft: eine Initiative mit Wirkung (interview)
VolkswagenStiftung. 15 July 2025.
Links: https://www.volkswagenstiftung.de/de/news/interview/ki-gesellschaft-eine-initiative-mit-wirkung
[67] Hakemusten arviointi; Tukea tieteelle ja taiteelle: lokakuun haku käynnistyy
Suomen Kulttuurirahasto (Finnish Cultural Foundation). 2025.
Links: https://skr.fi/apurahat/hakemusten-arviointi/ · https://skr.fi/ajankohtaista/tukea-tieteelle-ja-taiteelle-lokakuun-haku-kaynnistyy/
[68] Om søknadsprosessen og oppfølging av mottakerne
Sparebankstiftelsen DNB.
Links: https://sparebankstiftelsen.no/om-oss/om-soknadsprosessen-og-oppfolging-av-mottakerne/
[69] Personvernerklæring
SpareBank 1 Stiftelsen BV.
Links: https://www.sparebank1stiftelsenbv.no/personvernerklering
[70] Retningslinjer, vi følger (AI principles)
Villum Fonden.
Links: https://villumfonden.dk/da/artikel/retningslinjer-vi-foelger
[71] Neue Regeln für den Einsatz von Künstlicher Intelligenz bei Gutachten
Schweizerischer Nationalfonds (Swiss National Science Foundation). 20 August 2026.
Links: https://www.snf.ch/de/iOg3Y6tvhQO5yksA/news/ki-regeln-fuer-gutachtende
[72] Uitleg over AI
Fonds voor Cultuurparticipatie. Updated 1 March 2026.
Links: https://www.cultuurparticipatie.nl/beleid/uitleg-over-ai/
[73] Usein kysyttyjä kysymyksiä
Koneen Säätiö (Kone Foundation).
Links: https://koneensaatio.fi/apurahat/apurahanhakijalle/usein-kysyttyja-kysymyksia/
[74] Handbook on Data Protection in Humanitarian Action, Third Edition
International Committee of the Red Cross. 2024.
Links: https://www.icrc.org/en/data-protection-humanitarian-action-handbook
[75] Every meal counts: How WFP is using AI to reach more people, faster
World Food Programme. 19 May 2026.
Links: https://www.wfp.org/stories/every-meal-counts-how-wfp-using-ai-reach-more-people-faster
[76] RFQ: AI functionality for individual productivity in an NGO context
Norwegian Refugee Council. 10 April 2026.
Links: https://www.nrc.no/globalassets/pdf/tenders/ho/ai-for-personal-productivity/ai-for-personal-productivity-2026.pdf
[77] Personal Data and Privacy Policy
Norwegian Refugee Council. 23 May 2024, revised September 2024.
Links: https://www.nrc.no/globalassets/pdf/privacy/policies/english-personal-data-and-privacy-policy--2024.pdf
[78] UN shared Rohingya data without informed consent
Human Rights Watch. 15 June 2021.
Links: https://www.hrw.org/news/2021/06/15/un-shared-rohingya-data-without-informed-consent
[79] Guidance for using artificial intelligence in fundraising
Fundraising Regulator (United Kingdom). Updated 8 December 2025.
Links: https://www.fundraisingregulator.org.uk/about-fundraising/resources/guidance-using-artificial-intelligence-fundraising
[80] AI Readiness Playbook for Funders
Google.org. 2026.
Links: https://google.org/ai-playbook-for-funders
[81] Introducing an AI Governance Framework for Nonprofits
Microsoft Tech for Social Impact. 12 August 2024.
Links: https://techcommunity.microsoft.com/blog/nonprofittechies/introducing-an-ai-governance-framework-for-nonprofits/4217132
[82] Getting started with Claude for Nonprofits; OpenAI for Nonprofits
Anthropic and OpenAI help pages, as at November 2026.
Links: https://support.claude.com/en/articles/12893767-getting-started-with-claude-for-nonprofits · https://help.openai.com/en/articles/9359041-chatgpt-for-nonprofits
[83] Opening of the AI Centre at the Ministry of Awqaf
Al-Watan (Egypt) and Amwal Al-Ghad. 2 November 2026.
Links: https://elwatannews.com/news/details/8388832
[84] Information system development project for the Egyptian Awqaf Authority
Egyptian General Survey Authority.
Links: https://www.esa.gov.eg/awqaf.aspx
[85] پای هوش مصنوعی به مدیریت موقوفات باز شد (AI opens the door to Waqf management)
Tasnim News Agency (Iran). 20 October 2026 (28 Mehr 1405). Interview with Seyed Heydar Mortazavi, Director General of Information Technology and Communications, Awqaf and Charity Affairs Organisation. An earlier report of 3 October 2026 was cited by the research tools but not opened.
Links: https://www.tasnimnews.ir/fa/news/1405/07/28/3727478
[86] 2020 Yılı Performans Programı
T.C. Vakıflar Genel Müdürlüğü (General Directorate of Foundations, Türkiye).
Links: https://cdn.vgm.gov.tr/genelicerik/genelicerik_218_260519/2020-yili-performans-prog.pdf
[87] Directorate of Awqaf: 2024 achievements
Ministry of Habous and Islamic Affairs, Morocco. Page located by the research tools; the site could not be opened from this review's tools.
Links: https://habous.gov.ma/%D9%82%D8%A7%D8%A6%D9%85%D8%A9/%D8%A7%D9%85%D8%A7%D8%B1%D8%A9-%D8%A7%D9%84%D9%85%D8%A4%D9%85%D9%86%D9%8A%D9%86-2/21907-%D8%A7%D9%84%D9%86%D8%B8%D8%A7%D9%85-%D8%A7%D9%84%D9%85%D8%B9%D9%84%D9%88%D9%85%D8%A7%D8%AA%D9%8A-%D9%88%D8%A7%D9%84%D8%B1%D9%82%D9%85%D9%86%D8%A9-%D8%B3%D9%86%D8%A9-2024.html
[88] Waqf Funds Development Department announcement, 7 September 2026; Strategic Plan 2024
Ministry of Awqaf, Islamic Affairs and Holy Places, Jordan. The news item reports the launch of the first phase of the department's electronic services.
Links: https://awqaf.gov.jo/ar/NewsDetails/%D8%A7%D9%84%D8%AE%D9%84%D8%A7%D9%8A%D9%84%D8%A9_%D9%8A%D9%83%D8%B4%D9%81_%D8%B9%D9%86_%D8%AE%D8%B7%D8%A9_%D9%84%D9%85%D8%B6%D8%A7%D8%B9%D9%81%D8%A9_%D8%A5%D9%8A%D8%B1%D8%A7%D8%AF%D8%A7%D8%AA_%D8%AF%D8%A7%D8%A6%D8%B1%D8%A9_%D8%AA%D9%86%D9%85%D9%8A%D8%A9_%D8%A3%D9%85%D9%88%D8%A7%D9%84_%D8%A7%D9%84%D8%A3%D9%88%D9%82%D8%A7%D9%81_%D8%AE%D9%84%D8%A7%D9%84_%D8%A7%D9%84%D8%A3%D8%B9%D9%88%D8%A7%D9%85_%D8%A7%D9%84%D8%AB%D9%84%D8%A7%D8%AB%D8%A9_%D8%A7%D9%84%D9%85%D9%82%D8%A8%D9%84%D8%A9 · https://www.awqaf.gov.jo/EBV4.0/Root_Storage/AR/%D8%A7%D9%84%D8%AE%D8%B7%D8%A9_%D8%A7%D9%84%D8%A7%D8%B3%D8%AA%D8%B1%D8%A7%D8%AA%D9%8A%D8%AC%D9%8A%D8%A9_2024.pdf
[89] Waqf asset management service
Ministry of Endowments and Religious Affairs, Oman.
Links: https://www.mara.gov.om/arabic/ServiceDetails.aspx?id=104
[90] Procurement notice for AI analysis of contact-centre calls (No. 260941005284)
General Authority for Awqaf, Saudi Arabia, via the Etimad portal. 14 September 2026. Reported by the research tools; the record sits behind a verification page and was not independently opened for this report.
Links: https://tenders.etimad.sa/Announcement/SupplierAnnouncementDetails?announcmentIdString=MMr%2A%40%40%2A%2ACkmjnE3we1HRP%2A%40%40%2A%2AmwHA%3D%3D
[91] Direct marketing guidance: can we use special category data?
Information Commissioner's Office (United Kingdom).
Links: https://ico.org.uk/for-organisations/direct-marketing-and-privacy-and-electronic-communications/direct-marketing-guidance/plan-direct-marketing/
[92] Artificial intelligence as strategic enabler for transformation
Simsim, M. IsDB Institute blog. 3 September 2026.
Links: https://blogs.isdbinstitute.org/isdb-ai-strategic-enabler-transformation/
[93] IsDB 2025 Annual Development Effectiveness Report
Islamic Development Bank. June 2026. "Introduction of an AI-enabled ADER Advisory Agent", corroborated on the IsDB ADER 2025 online summary.
Links: https://beta.isdb.org/sites/default/files/media/documents/2026-06/IsDB-ADER%20Report-final%2011%20June%202026_compressed.pdf
[94] IEvD hosts insightful knowledge session on Islamic Finance Artificial Intelligence Assistant (IFAA)
Islamic Development Bank, Independent Evaluation Department. 19 August 2024.
Links: https://www.isdb.org/independent-evaluation/ievd-hosts-insightful-knowledge-session-on-islamic-finance-artificial-intelligence-assistant-ifaa
[95] IsDB Institute unveils 2025 Annual Report during Group Annual Meetings in Baku
IsDB Institute. June 2026.
Links: https://isdbinstitute.org/isdbi-2025-annual-report/
[96] The IsDBI and Arab Monetary Fund (AMF) deliver training on using artificial intelligence to foster the Islamic financial industry
IsDB Institute, distributed by APO Group. 3 to 4 December 2025.
Links: https://isdbinstitute.org/isdbi-and-amf-training-on-using-ai/
[97] IsDBI announces winners of Applied AI in Islamic Finance Competency Challenge
IsDB Institute, distributed by APO Group. 18 June 2026.
Links: https://isdbinstitute.org/isdbi-applied-ai-islamic-finance-winners/
[98] Frequently asked questions by beneficiaries
Awqaf Properties Investment Fund, Islamic Development Bank.
Links: https://www.isdb.org/apif/faq
[99] World Development Report 2026: The promise of artificial intelligence
World Bank. August 2026. AI-use disclosure statement.
Links: https://ppp.worldbank.org/sites/default/files/2026-08/World%20Development%20Report%202026%20The%20Promise%20of%20Artificial%20Intelligence.pdf
[100] Digital ADB: An operational approach for digital transformation 2026-2030
Asian Development Bank. July 2026.
Links: https://www.adb.org/sites/default/files/institutional-document/1158681/digital-adb-operational-approach-2026-2030.pdf
[101] Evaluation approach paper: Asian Development Fund 13 and 14
Asian Development Bank, Independent Evaluation Department. August 2026. Paragraph 33 on the use of AI in the evaluation.
Links: https://www.adb.org/sites/default/files/Evaluation%20Document/1173626/files/APPROVED%20Approach%20Paper_ADF%2013-14_circ%202.pdf
[102] Creating impact through technology-enabled infrastructure
Asian Infrastructure Investment Bank. 2026. AI-use disclosure statement.
Links: https://www.aiib.org/en/publications/overview/_common/pdf/creating-impact-through-technology-enabled-infrastructure-2026.pdf
[103] AIIB Annual Report 2025
Asian Infrastructure Investment Bank. 2026. Internal AI platform statement confirmed by ChatGPT's review; the PDF could not be opened from Claude's tools.
Links: https://www.aiib.org/en/news-events/annual-report/2025/_common/pdf/2025-AIIB-Annual-Report.pdf
[104] Tech report: Generative AI
Latorre, L., Rego, E., Cerrato, I., Zarate, J. D. and De Leo, L. Inter-American Development Bank. March 2025. Section on projects, prototypes and pilots.
Links: https://doi.org/10.18235/0013473 · https://publications.iadb.org/publications/english/document/Tech-Report-Generative-AI.pdf
[105] Integrity and Anti-Corruption Report 2025
European Bank for Reconstruction and Development, Office of the Chief Compliance Officer. 2026. Reported by the research tools; not independently opened for this report.
Links: https://www.ebrd.com/content/dam/ebrd_dxp/assets/pdfs/office-of-chief-compliance-officer/integrity-and-anti-corruption-reports/ebrd-integrity-and-anti-corruption-report-2025.pdf
[106] Article on BAZNAS digital zakat management
IJAZ BAZNAS (Indonesian Journal of Zakat), BAZNAS. Reported by the research tools; not independently opened for this report.
Links: https://www.ijazbaznas.com/index.php/journal/article/download/769/281/2293