Permissible by Default

Ask what Islam says about artificial intelligence and you will mostly find two kinds of answer: a fatwa about one particular use, or a news story saying that some council has “called for ethics”. Both are real, and neither tells you what the institutions that speak for large parts of the Muslim world have actually put in writing. Those texts exist now. Since January 2023 an Islamic legal academy, a world conference of fatwa authorities, the Grand Imam of Al-Azhar, an American research institute and one of the most senior living jurists have each published on AI, in English, under their own names.

This essay reads them one statement at a time. It is the institutional companion to The Use, Not the Tool, which takes the single question “is using AI a sin?” and reads the individual fatwas on it, including the ones on AI pictures. It does not repeat that essay’s reading of those fatwas; where they matter here, it points to them.

How the sources were handled. Every text below was opened with a public GET on 10 October 2026 and quoted from the English the publisher itself put out. Several of these statements were first spoken or written in Arabic; where that is so, the English quoted is the version the institution published, and I say whose it is. I have not translated any Arabic text for this page. Every quoted sentence is recorded, with its address and the date it was read, in this site’s source ledger.

What they have published, by date

Two things stand out before any detail. First, nobody on this list says AI is forbidden, and two of them, the Fiqh Academy and Yaqeen, say in nearly the same words that it is permitted by default and conditioned by use. Second, only one of these is a ruling in the strict sense. The Academy’s resolution is the decision of the legal academy of the Organization of Islamic Cooperation, an intergovernmental body of 57 states, and the Dar al-Ifta text is a fatwa. The rest are speeches, a conference statement and an institute’s account of its own practice. They carry weight, but a different kind.

The ruling: the Fiqh Academy’s six conditions

The most formal text is the shortest. The International Islamic Fiqh Academy, based in Jeddah, met for its 26th session in Doha in May 2025, and its Resolution No. 258 on AI runs to a few hundred words. It says it reviewed the research papers submitted on the subject and ICESCO’s 2025 Riyadh Charter on AI for the Islamic World, and then rules in two steps. First it defines AI as a technology that simulates human intelligence and brings about numerous benefits, though it is not without potential harms. Then:

The default ruling regarding the development and use of artificial intelligence is permissibility (ibāhah), subject to the following ethical and legal conditions:International Islamic Fiqh Academy, Resolution No. 258 (3/26), Artificial Intelligence: Its Rulings, Guidelines, and Ethics, May 2025, Second

The six conditions are worth reading as a list, because they are the closest thing to an official Islamic checklist for AI that exists in English. The purpose of making, using and funding it, and its outcomes, must be lawful. It must bring benefit and prevent harm. It must not insult or abuse beliefs, religions, or religious symbols. It must protect information and preserve public and private rights. It must not threaten individual, social or national security. And it must adhere to honesty, proper documentation, and transparency in its use.

Notice the word development. The ruling covers the people who build AI, not only those who use it, and its first condition reaches as far as funding. Notice also what the Academy left open. Its first recommendation is Conducting a study on the legal ruling regarding granting AI legal personality. In other words, the most authoritative body on this list has not decided whether an AI could be a party in law, and has said so in writing.

The fatwa authorities: AI for muftis, not as a mufti

Egypt’s Dar al-Ifta has published on AI from two directions, and reading them together removes what can look like a contradiction.

Its December 2025 fatwa, read in detail in The Use, Not the Tool, forbids relying on AI applications to obtain a fatwa. Four months earlier, in August 2025, Dar al-Ifta hosted the General Secretariat for Fatwa Authorities Worldwide, whose conference drew participants from more than eighty countries on the theme of preparing muftis for the age of AI. Its closing statement is anything but hostile to the technology. Its sixth recommendation calls on scholars to integrate artificial intelligence technologies into research and daʿwah; its thirteenth calls for Developing artificial intelligence tools to serve moderate fatwa, with apps and platforms in many languages; its eighth wants AI skills taught in the training of muftis.

The fifteenth recommendation is where the two texts meet. It asks for public campaigns about the dangers of relying on unverified fatwas from open platforms, and seeking Sharia answers from artificial intelligence applications that are not subject to review and evaluation. So the line is not between AI and no AI. It runs between an answer a qualified scholar or an accredited institution stands behind and an answer nobody stands behind. The same week, the Muslim Council of Elders reported that the Grand Imam of Al-Azhar, receiving the conference’s delegates, had prayed that God guide scholars contributing to the use of artificial intelligence applications to serve muftis and facilitate their work, in its release of 14 August 2025.

The Yaqeen Institute, a research institute based in Texas with no state behind it, draws the same line in its own practice. Its AI assistant, AQSA, searches the institute’s published research, and the institute is explicit about its limit: AQSA does not provide rulings or personal religious advice. Later the essay says it in five words: It is not an AI mufti. Its reason is a story about the Companion ʿAbdullah ibn ʿAbbas, who gave two men opposite answers to the same question because he judged what each intended to do. The point is that a ruling depends on the person asking, and the essay’s comment is: Imagine outsourcing this to a machine—such nuance and intuition would be lost entirely.

Al-Azhar: who will regulate it, and who answers for it

Al-Azhar’s Grand Imam, Dr. Ahmed Al-Tayeb, has not issued a ruling on AI. What he has said comes through the English releases of the Muslim Council of Elders, which he chairs, and the clearest of them is from June 2026, when he received the UN Deputy Secretary-General, Amina Mohammed. The release reports him warning that excessive reliance on AI is already harming critical thinking and creativity and has become a way of appropriating the work of others, and quotes him directly:

The question that should concern the world today is not how far technology and artificial intelligence will advance, but rather who will regulate this advancement, on what system of values and ethics it will be based, and who will be held accountable.Muslim Council of Elders, English release of 28 June 2026, quoting Dr. Ahmed Al-Tayeb, Grand Imam of Al-Azhar

That is the English the Council published. He was presumably speaking Arabic, and I have not seen an Arabic transcript, so the wording is the Council’s rendering of what he said. The worry it states is about power more than doctrine. The same release has him concerned about AI being controlled by a limited few capable of influencing the fate of nations, which is close to Pope Leo XIV’s line in Magnifica Humanitas that a more moral AI is not enough if that morality is determined by a few, read in A Pale Reflection.

Sheikh bin Bayyah at the Vatican: the Rome Call in context

The oldest texts on this list come from the Vatican, on 10 January 2023, when Jewish and Muslim leaders signed the Rome Call for AI Ethics. The Muslim side was the Abu Dhabi Forum for Peace, whose president, Sheikh Abdallah bin Bayyah, also chairs the UAE Fatwa Council. A Pale Reflection records the signature but, as it says, did not read his speech. Here it is.

He begins with welcome, not alarm. Islam, he says, encourages the pursuit of knowledge, and quotes the Qur’anic prayer Say, ‘My Lord, increase me in knowledge.’ Religion, in his account, was not a hindrance to innovation, but rather a stimulus that motivated creativity and achievement. Then he lists six questions AI raises, and they are sharper than most of what a press release carries. The first is Who is responsible for the mistakes of artificial intelligence? The fifth is How do we control a system that may prove smarter than we are? No other text on this list puts that question.

His framework is the one part of the speech that is distinctively Islamic, and it gives the clearest picture of how a jurist reasons about a new technology:

These dangers could pose a threat to humanity from the perspective of what we Muslims term the five protected universals: religion, life, reason, property, and human integrity, the last of which includes family and individual human dignity.Sheikh Abdallah bin Bayyah, speech at AI Ethics: An Abrahamic commitment to the Rome Call, Vatican City, 10 January 2023

He maps the dangers onto them one by one: life, through AI as an uncontrolled weapon; human integrity, through lost privacy; property, through electronic theft and piracy. Beyond the five, he warns of division in societies through false information and the swaying of voters. His legal anchor is a saying of the Prophet that he quotes as There should be no harm nor reciprocation of harm. And he gives an image for what ethics does to a technology, which is not a ban but steering: a framework that works like a brake which protects cars from collision and destruction, or as a compass that guides ships in the midst of a vast ocean.

Read the signature in that light. The Joint Declaration the three organisations issued that day is blunt in one respect: it says plainly that voluntary regulation is not enough and asks for international regulation, and it asks governments to refrain from weaponising AI. But its six principles (transparency, inclusion, responsibility, impartiality, reliability, security and privacy) are general, and the Islamic reasoning in it is thin. That reasoning is in bin Bayyah’s speech, not in what he signed. A separate track exists too: in May 2025 the Muslim Council of Elders reported that the Grand Imam of Al-Azhar had initiated, together with the late Pope Francis, an international document on the ethics of artificial intelligence, and that discussions with the Vatican’s new leadership were ongoing (release of 27 May 2025). As of 10 October 2026 I have not found that document published.

The questions the sources raise

AI-generated fatwas and rulings. Every source that addresses it agrees: AI may help a scholar and may not replace one. Dar al-Ifta’s fatwa forbids relying on AI for a fatwa; the fatwa conference wants AI tools for muftis but warns the public off unreviewed AI answers; Yaqeen’s assistant does not rule. Where they differ is on how hopeful to be. The conference wants fatwa apps in many languages; Yaqeen sees the limit as built into what a ruling is.

Image-making. None of the institutional texts here rules on whether an AI picture of a living being is forbidden image-making; the fatwas that do, and that disagree, are read in The Use, Not the Tool. What the institutional texts give is practice. Yaqeen states that it does not use AI to create realistic depictions of living people or imitate their voices (i.e., deepfakes), and limits AI visuals to things like background footage. The Fiqh Academy’s third condition, that AI must not insult beliefs or religious symbols, would cover a different kind of image again.

Authorship. Yaqeen is the most direct: Authorship is not just about producing material; it reflects intention and accountability before Allah. Its authors affirm on submission that the content and conclusions are theirs. The Academy’s sixth condition asks for honesty, documentation and transparency in use, and the Grand Imam’s June 2026 warning names AI as a means of appropriating the work of others. None of them forbids using AI in writing; all of them place the responsibility for the words on a named person.

Ethics and power. The Joint Declaration says voluntary regulation is not enough, and the Grand Imam asks who will regulate AI and on what values. Bin Bayyah adds that laws and regulators must set standards to prevent harm. The Fiqh Academy, a legal body, speaks instead to the conscience of the individual Muslim: its conditions bind the believer whatever any government does.

Where they disagree

The sources agree on the starting point, permitted unless the use is wrong, and on the end point, a human answers for it. They differ in three places.

How far to go with AI in religious knowledge. The fatwa conference is the most enthusiastic text here; it wants AI built into training, research and the delivery of fatwas. Yaqeen uses AI too, but its essay is more wary of what it calls AI slop and of a model that can misquote a hadith or strip a ruling of its context while sounding eloquent and confident. The difference is of emphasis, not doctrine, but it decides what gets built.

Whether AI is only a tool. Yaqeen treats it as one throughout: AI does not make decisions or carry moral weight. Bin Bayyah’s fifth question, about controlling a system smarter than us, is not a question anyone asks about a hammer. And the Fiqh Academy has formally left open whether AI could have legal personality. These are not contradictions, since a thing can be a tool and still hard to control, but they do not sit together as easily as the shared word “permissible” suggests.

Who speaks for Islam on it. The fatwa conference says that official fatwa institutions are the fortified shield protecting Sharia from erroneous fatwas. Yaqeen is not an official fatwa institution and reaches its own position on its own scholarly authority. On AI the two agree; on who gets to say so, they start from different places.

The strongest objection: this is official Islam talking about itself

Here is the case against taking this list as “what Islam says about AI”, made as strongly as I can.

Nearly every voice on it is official, and most are close to a state. The fatwa conference met under the patronage of Egypt’s president and thanks him at length. The Fiqh Academy is an organ of an intergovernmental body. Sheikh bin Bayyah spoke for a forum based in Abu Dhabi and closed his speech by praising the UAE’s ambition to lead in AI. Read that way, the one firm ruling, that you may not rely on AI for a fatwa, protects the business of the people issuing it: a chatbot that answers religious questions competes directly with a fatwa hotline. And the ethics on offer, transparency, privacy, reliability, no bias, are the same list any technology company publishes. Put “Islamic” in front of a generic AI-principles document and you have not learned anything about Islam.

Much of that is fair. These texts do come from the top, and they cannot tell you what an imam in Lahore or a student in Leeds actually thinks. The overlap with secular principles is real, and the Joint Declaration in particular could have been written by anyone.

But the objection proves less than it claims. The ruling on AI fatwas is not just an institution guarding its patch: its reason, that a ruling depends on the circumstances and intention of the person asking, is the same reason Yaqeen gives with the story of Ibn ʿAbbas, and Yaqeen has no state and no hotline to protect. And the reasoning is not generic. “Permissible by default” is a specific legal principle with a long history in Islamic jurisprudence, the maxim that the means take the ruling of the ends is a specific inference, and bin Bayyah’s five protected universals are a specific way of sorting harms that a secular ethics board would not produce. The conclusions often match secular ones. The reasons are different, and the reasons are what tell you how these bodies will rule on the next question.

Who is missing

This essay covers only texts I could open and read in the publisher’s own English. That leaves real gaps, and I would rather name them than fill them from news reports.

• The Muslim World League. Its Secretary-General, Dr. Mohammad Al-Issa, has spoken on AI ethics in Fez and Kuala Lumpur, but I found only press coverage of those speeches, not a text on the League’s own site, so I quote neither.
• ICESCO’s Riyadh Charter on AI for the Islamic World, which the Fiqh Academy says it reviewed. I did not open its text.
• The UAE Fatwa Council and Dubai’s fatwa department. Sheikh bin Bayyah chairs the first, but I found no ruling of its own on AI; the Dubai department’s English news page on using AI in its fatwa system refused my request (HTTP 403).
• Shia authorities. I found no published English text on AI from the offices of the senior Shia jurists, only news reports and an academic roundtable. Every source in this essay is Sunni.
• South and Southeast Asia. Indonesia, Pakistan, India and Bangladesh are home to a large share of the world’s Muslims. I opened no ruling on AI from their councils or seminaries.
• Arabic-only texts. Some of what these bodies have published exists only in Arabic. I have not quoted any of it, because I would have had to translate it myself.

That is what this list covers. It is not a census of what Muslims think.

What this leaves

On the evidence opened here, the institutions that have spoken agree on more than the headlines suggest. AI is permitted by default and judged by what it is used for. A person, not a program, answers for a ruling, a text or a harm. Ethics needs law behind it, not just company promises. They part on how hopeful to be about AI in religious knowledge, on whether “tool” is the whole story, and on who has the standing to say so.

For what the churches have published, document by document, read A Pale Reflection; for the individual fatwas, including those on AI images, and for Christian and Jewish rulings on using AI, read The Use, Not the Tool. Books by believers who have written on AI at length are on the reading list.