For much of Islamic history, religious authority has depended on human judgment: a scholar, trained in jurisprudence and accountable to a school of thought, responded to a question shaped by context, intention, and circumstance. That arrangement was never perfect or uncontested, but it was grounded in a tradition and relationship—between petitioner (mustafti) and jurist, text and exegesis, ethics and responsibility. That relationship is now markedly less stable. Artificial intelligence, usually framed in tech and policy discourse as an objective instrument of efficiency, is increasingly being folded into the production of Islamic knowledge. What’s happening does not resemble a clean break so much as a quiet reordering of how authority, trust, and legitimacy take shape.
In 2019, Dubai launched what was described as the world’s first “Virtual Ifta” program, marking an early attempt to digitize ifta—the practice, particularly in the Sunni context, of issuing nonbinding legal opinions (fatwas) in response to religious questions. The system introduced machine learning into a domain historically grounded in human interpretation, signaling a shift toward faster, more scalable forms of religious guidance.
By August 2025, this experiment had expanded into a broader national strategy. In collaboration with the UAE Council for Fatwa and a network of international scholars, the UAE announced plans to integrate generative AI into its religious infrastructure. The Virtual Ifta Programme is designed to process religious queries and generate automated responses once a defined confidence threshold is met, while maintaining formal oversight by authorized scholars.
Framed as a way to modernize access and limit unauthorized or extremist rulings, the initiative positions AI as both a tool of efficiency and a mechanism of institutional control. The language surrounding these projects is careful. AI is framed as a support for scholars, not a substitute. Human oversight is stressed again and again. But even amid those assurances, something else is taking shape. Religious interpretation is being translated into datasets, probabilities, and outputs, recast as something that can be standardized, scaled, and delivered instantly. What begins to erode is not the availability of guidance, but the relational and deliberative conditions that once gave that guidance meaning.
Away from official statements and policy forums, this shift surfaces differently. On TikTok and YouTube, English language scholars, muftis, and religious educators find themselves responding to the same questions: Can AI be trusted for religious guidance? Can it issue fatwas?
In a widely circulated clip, Mufti Monawwar Ateeq cautioned that while AI may assist with gathering information, summarizing opinions, and conducting preliminary research, it cannot replace traditional scholars. He emphasized that such systems require human verification and scholarly oversight, noting that they lack empathy, contextual judgment, and the depth of insight necessary for religious guidance. Certain rulings in Islamic law, he notes, hinge distinctions between diyanah (one’s accountability before God) and qada (formal legal judgement); these distinctions require human discernment. Similar concerns surface elsewhere online. Responding to a question submitted to his online platform, Assim Al-Hakeem, a well-known Saudi Sheikh, cautioned that AI-generated religious answers are dangerous and unreliable and cannot substitute for qualified scholarly judgment.
A similar concern appears across more informal digital spaces, like Instagram. These interventions function less as sermons than as boundary-setting responses to a question that keeps resurfacing—scholars are not introducing the topic, but they must respond to it. The question itself signals that algorithmic tools have already entered the space where authority is being negotiated.
The AI-assisted fatwa system and the short-form video response point to a shift in how religious authority is encountered. In both cases, authority appears through interfaces rather than sustained relationships, shaped by systems that prioritize speed and reach over deliberation. The state-backed platform and the online religious figure differ in form, yet they operate under similar conditions. Immediacy tends to be rewarded, while hesitation circulates less easily. Accessibility begins to stand in for credibility, and context becomes harder to maintain.
This matters because fatwas were never meant to function as universal answers. A fatwa is offered in response to a particular situation, shaped as much by circumstance as by text. Scholars like Sherman Jackson have long emphasized that Islamic authority does not produce certainty on demand, but instead, functions in the context of moral reasoning—weighing competing values, recognizing limits, and remaining accountable to a community. Interpretation is not an obstacle to overcome. It is part of the work’s ethical and legal substance.
Recent scholarship on AI and Islamic law points to the difficulty of translating this tradition into computational systems. In a roundtable discussion held in 2025, Intisar Rabb and Mairaj Sayed noted that current large language models struggle precisely where Islamic legal reasoning is most demanding: in context, nuance, and the interpretation of complex source traditions.
AI unsettles that arrangement by changing what judgment is formed. Machine-learning systems generate patterned outputs instead of engaging in deliberation. Even when institutions insist that AI merely assists scholars, its presence shifts expectations. Why wait for interpretation when an answer arrives instantly? Why live with ambiguity when clarity is optimized?
These frictions show up in small but telling ways. Even on early internet fatwa portals such as IslamQA, framed religious answers around deference to qualified scholars; in its guidance on AI-generated fatwas, the site warns users not to rely on artificial intelligence and to consult trustworthy scholars instead. The language hedges authority even as the interface promises immediacy and scale. At the state level, meanwhile, initiatives such as UAE’s AI-assisted fatwa program expand automated infrastructures while repeating assurances about human supervision. None of this reads as outright rejection. It reads more like containment. Lines are drawn, tested, adjusted, and drawn again.
What begins to emerge is not a dialectical relationship between two reasoning agents but a reconfigured field of authority shaped by their interaction. AI systems are trained on religious texts and scholarly traditions, but once they circulate, they reshape how guidance is sought. Authority is redistributed across new technical systems that reorganize how religious guidance is assessed and trusted.
The political stakes of that shift become sharper in state contexts. In the Gulf, AI-mediated religious initiatives sit comfortably alongside broader technocratic governance projects, from national AI strategies to state-run digital services. In the case of the United Arab Emirates, the General Authority of Islamic Affairs and Endowments has launched programs to train thousands of employees in artificial intelligence as part of a wider national push toward digital transformation. Officials describe AI as a tool that can facilitate access to religious information, reduce response time, and strengthen what they call the country’s “intellectual and informational resilience.”
These efforts mirror UAE’s wider investment in “smart government” infrastructure, where automation, data integration, and predictive systems are framed as tools of stability and modernization. Within this policy environment, religious authority is increasingly presented as something that can be optimized through efficiency, safeguarded through institutional control, and justified through the language of moderation and security. Innovation becomes a governing style. Interpretation becomes policy to administer.
For Muslims in the diaspora, the picture looks different but no less charged. Many already encounter Islam primarily through apps, feeds, and short-form content rather than through local institutions. These encounters unfold under conditions of surveillance and data extraction, where visibility is both unavoidable and risky. AI intensifies this terrain. It doesn’t just mediate religious practice; it shapes how Muslims are rendered legible to platforms, states, and other monitoring publics. From AI-generated Islamophobic imagery circulating in India to algorithmic policing in Western contexts, the same systems that promise guidance also participate in regimes of suspicion and control.
Seen this way, AI’s role in Islamic life cannot be separated from broader structures of power. Religious mediation, platform governance, and racialized surveillance fold into one another. Authority is never neutral, and neither is the technology that carries it.
The question, then, is not whether AI belongs in religious life—it is already there—but how authority will be negotiated in its presence. What happens when guidance can be produced at scale and delivered instantly but becomes impossible to hold accountable? What kinds of Muslim futures are being imagined when interpretation is compressed into a response window? When trust is delegated to systems that cannot be questioned in return?
We may be living through a moment before these arrangements fully settle. The stakes are not primarily doctrinal. They are also political and cultural. They concern who gets to interpret, who gets to decide, and who bears responsibility when judgment is mediated by machines—shifting authority away from locally accountable scholars toward opaque technical systems, state institutions, and platforms that cannot be easily questioned or contested. The implications lie in how these systems are implemented and governed—through decisions that are often hidden and framed as progress.
