Authored By: Sumera Sabir
Graduate from The University of Law
I. Introduction
The rise of Artificial Intelligence (AI) has become more widespread across a range of professional fields, including regulatory decision-making procedures, and legal and commercial contexts.1 AI refers to the imitation of human intelligence by a machine or system.2 The aim of AI is to devise a machine that is able to imitate human behaviour and think like humans, achieved through learning, planning, perceiving, predicting, and reasoning.3 It has also been more frequently employed within religious and legal judgements, including Islamic legal (Shariah) advising and in computerised fatwa platforms.4
The increased utilisation of such technologies has prompted questions amongst scholars. These questions concern whether AI can legitimately execute the role of a mufti and whether it is able to exercise independent legal reasoning (ijtihād).5
This article assesses the limitations of AI issuing fatwas and proceeds as follows. Section II outlines the Shariah framework for the analysis. Section III examines some of the prime limitations of AI in Shariah law. Section IV proposes solutions to the limitations identified in Section III.
II. The Legal Framework
A. Revealed and Derived Sources
The primary legal framework for Shariah law is derived from two categories of sources: revealed and derived.6
The revealed sources, namely the Qur’an and the Sunnah (Hadith), are considered the primary and most definitive foundations of Islamic law. The Qur’an is regarded as God’s word, and the Hadith as Prophetic tradition. These serve as the groundwork for consensus and legal authority in Islamic jurisprudence.7
Derived sources emerge from human reasoning through ijtihad; this occurs when the revealed sources offer no explicit ruling, and Muslim jurists must formulate rules based on independent reasoning.8 There are specific requirements that must be met to be qualified to perform ijtihad.9
B. Fatwa
A fatwa is an authoritative answer provided by an Islamic legal expert to a specific religious issue, accordingly establishing a particular hukm (rule or decision) for its implementation.10 The Qur’an and Hadith serve as the basis when a fatwa is issued.11
A fatwa issuer is called a mufti (mujtahid, faqih, alim). An individual who needs a religious solution to a problem either extracts it from the sources (if one has the relevant knowledge, one is a mujtahid) or, if one cannot do so, refers to competent people in this area.12
Issuing a fatwa is a highly responsible and serious protocol, and it must be executed based on the primary sources (Qur’an and Hadith) of Shariah law. The mufti must meet extremely high standards, since fatwas with errors can cause disagreements and discord amongst Muslims.13
C. Ijtihad
Since the lifetime of the Prophet (peace and blessings be upon him), if one could not find an express legal ruling in the Qur’an or Sunnah, then ijtihad was exercised, so that one had the ability to arrive at a conclusion independently. The Prophet (peace and blessings upon him) would then inform them who was right and who was incorrect.14 AI does not have the ability to perform ijtihad, as it does not have the capacity to arrive at a conclusion independently.
III. Limitations of AI
A. The Absence of Human Qualities
The major limitation posed by AI is the absence of human qualities. These qualities are among the most vital components in issuing fatwas. Human judgement is not simply technical; it requires ethical judgement, awareness of context, and an understanding of spirituality. A study found that human-centred qualities are vital for producing rulings that are not only legally valid but also account for ethical and spiritual factors.15 Since AI is the simulation of human intelligence, it is “devoid of conscience, moral agency and intentionality.”16 AI cannot genuinely replicate human reasoning, as it operates on statistical correlation rather than intentional reasoning.17 This is evidence of the shortcomings AI would have if it issued fatwas and replaced the role of a mufti. If the judgement issued was erroneous, then it would constitute negligence.
B. Hallucinations and Errors
There have been substantial rates of factual hallucinations and errors in recent research evaluating large language models (LLMs) in Islamic law, raising concerns about their validity for issuing Islamic guidance.18 One study found:“The best system reaches 67.65% correctness with 21.25% hallucination, while several models fall below 35% correctness and exceed 55% hallucination.”19
These statistics demonstrate that AI-generated Islamic legal rulings can produce numerous fabricated outputs. These statistics establish that there is a fundamental need for rigorous regulations and supervision in Islamic law; LLMs are not presently competent to generate adequate outputs.
IV. Proposed Solutions
The article proposes the following three solutions for the limitations that have been outlined in the preceding section.
A. Ongoing Discussions
First, ongoing discussions in this area are mandatory between Islamic jurists, legal theorists, and AI regulators. Since AI is increasingly on the rise, modifications to the technology will continue to transpire, so discussions among these specialists will continue. There is a visible necessity for formal guidelines and robust compliance, to ensure that if any predictable or unpredictable issues arise, preventative measures are already in place. Since a substantial number of Muslims are utilising AI for legal advice, there is a need for ongoing conversations about how to regulate its use ethically.
B. Human-in-the-Loop
Second, the article proposes the adoption of a Human-in-the-Loop (HITL) model. This is a framework in which humans must approve and review every specific decision.20 This would enable individuals who are experts in the field of Islamic law to oversee the information that is being shared. This ensures that the power remains with the mufti, who remains the actual issuer of the fatwa, fulfilling their spiritual and legal duty. It also protects the interests of Muslims who are using AI in matters of Islamic law, and would ensure that robust regulations are in place to address current and future challenges.
C. Informational Content Disclaimer
Third, there is a need for an informational content disclaimer. The disclaimer should declare that any AI information on fatwas provided should not be substituted for professional Islamic legal advice from a mufti, and that its use is for informational or knowledge purposes only. This disclaimer safeguards Muslims’ interests by prompting them to consult with Islamic experts and to urge them to employ the information solely for knowledge or informational reasons.
V. Conclusion
The article has presented significant limitations that would arise if AI were to issue Islamic legal rulings independently. The legal framework in Islamic law would be undermined without supervision, which is imperative to prevent this from occurring. This article has argued that AI cannot deliver independent judgments because it lacks the capacity to engage in ijtihād, which requires human reasoning. Therefore, it would not be able to issue fatwas independently.
The article proposes that the solutions outlined in Section IV be adopted. These include maintaining ongoing conversations amongst specialists (Islamic jurists, legal theorists, and AI regulators), and adopting a HITL model that retains decision-making power with the relevant individuals in the field of Islamic law.
This is not to say that AI should be wholly disregarded for use in Islamic law. It is advantageous as an auxiliary tool for rapid information retrieval, legal research, and systematic analysis, and would therefore save jurists time and money when issuing fatwas. However, the positioning needs to make clear that AI is not a decision-making tool but an auxiliary instrument.
The article argues that, since AI imitates human intelligence, it can produce errors that can lead to fabrications in religious rulings. There is a risk that this could alter the true meaning of the scriptures in Islamic law, whether from the revealed or derived sources. This results in failing to uphold an individual’s rights equitably, making it contrary to the teachings of Islam.
Notes
1. European Parliament, Artificial Intelligence: Potential Benefits and Ethical Considerations (2020) 1.
2. Y Xu and others, ‘Artificial Intelligence: A Powerful Paradigm for Scientific Research’ (2021) 2(4) The Innovation 100179, 1.
3. Y Xu and others, ‘Artificial Intelligence: A Powerful Paradigm for Scientific Research’ (2021) 2(4) The Innovation 100179, 1.
4. Alwy Ahmed and others, “Can Artificial Intelligence Perform Ijtihad? Authority, Epistemology, and Legitimacy in Digital Islamic Jurisprudence” (2026) 38 Suhuf 172, 172 https://doi.org/10.23917/suhuf.v38i1.16437.
5. Ibid 173.
6. Taha Jabir Al-Alwani, Anas S Al-Shaikh-Ali and Yusuf Talal DeLorenzo, Source Methodology in Islamic Jurisprudence: (Usul al-Fiqh al-Islami) (International Institute of Islamic Thought (IIIT) 2003) 4.
7. Ibid 4–5.
8. Taha Jabir Al-Alwani, Anas S Al-Shaikh-Ali and Yusuf Talal DeLorenzo, Source Methodology in Islamic Jurisprudence: (Usul al-Fiqh al-Islami) (International Institute of Islamic Thought (IIIT) 2003).
9. Ibid 1.
10. Hussein Idrisov and Rizvan Ibragimov, “Fatwa Issuance in the Islamic Law: The Specific of Implementation under the Conditions of the Chechen Republic (History and Modernity)” (2023) 172 SHS Web of Conferences 3 http://dx.doi.org/10.1051/shsconf/202317204001.
11. Ibid 1.
12. Hussein Idrisov and Rizvan Ibragimov, “Fatwa Issuance in the Islamic Law: The Specific of Implementation under the Conditions of the Chechen Republic (History and Modernity)” (2023) 172 SHS Web of Conferences 2–3 http://dx.doi.org/10.1051/shsconf/202317204001.
13. Hussein Idrisov and Rizvan Ibragimov, “Fatwa Issuance in the Islamic Law: The Specific of Implementation under the Conditions of the Chechen Republic (History and Modernity)” (2023) 172 SHS Web of Conferences 6 http://dx.doi.org/10.1051/shsconf/202317204001.
14. Taha Jabir Al-Alwani, Anas S. Al-Shaikh-Ali and Yusuf Talal DeLorenzo, Source Methodology in Islamic Jurisprudence (Uṣūl al-Fiqh al-Islāmī) (International Institute of Islamic Thought 2003) 5–6.
15. Mirismoil Usmonov, “From Human Scholars to AI Fatwas: Media, Ethics, and the Limits of AI in Islamic Religious Communication” (2025) 5 Al-i’lam – Journal of Contemporary Islamic Communication and Media 75 https://doi.org/10.33102/jcicom.vol5no1.125.
16. Alwy Ahmed and others, “Can Artificial Intelligence Perform Ijtihad? Authority, Epistemology, and Legitimacy in Digital Islamic Jurisprudence” (2026) 38 Suhuf 172, 173 https://doi.org/10.23917/suhuf.v38i1.16437.
17. Mohammed Gamal Abdelnour, “Artificial Intelligence and the Islamic Theology of Technology: From ‘Means’ to ‘Meanings’ and from ‘Minds’ to ‘Hearts'” (2025) 16 Religions 796, 4 https://doi.org/10.3390/rel16060796.
18. Ezieddin Elmahjub and others, ‘IslamicLegalBench: Evaluating LLMs Knowledge and Reasoning of Islamic Law Across 1,200 Years of Islamic Pluralist Legal Traditions’ (2026) arXiv Working Paper.
19. Ezieddin Elmahjub and others, “IslamicLegalBench: Evaluating LLMs Knowledge and Reasoning of Islamic Law across 1,200 Years of Islamic Pluralist Legal Traditions” [2026] Open MIND 30.
20. Mirismoil Usmonov, “From Human Scholars to AI Fatwas: Media, Ethics, and the Limits of AI in Islamic Religious Communication” (2025) 5 Al-i’lam – Journal of Contemporary Islamic Communication and Media 75–76 https://doi.org/10.33102/jcicom.vol5no1.125.
Reference(S):
Books
Al-Alwani TJ, Al-Shaikh-Ali Anas S and DeLorenzo YT, Source Methodology in Islamic Jurisprudence: (Usul al-Fiqh al-Islami) (International Institute of Islamic Thought (IIIT) 2003)
Secondary Sources
Ahmed A and others, “Can Artificial Intelligence Perform Ijtihad? Authority, Epistemology, and Legitimacy in Digital Islamic Jurisprudence” (2026) 38 Suhuf https://doi.org/10.23917/suhuf.v38i1.16437
Abdelnour MG, “Artificial Intelligence and the Islamic Theology of Technology: From ‘Means’ to ‘Meanings’ and from ‘Minds’ to ‘Hearts'” (2025) 16 Religions 796 https://doi.org/10.3390/rel16060796
Billah MM and Zahid Z, “Artificial Intelligence in Islamic Fatwa and Shariah Advisory: Opportunities, Limitations, and the Imperative of Human-in-the-Loop Shariah Governance” (2026) 03 International Journal of Innovative Inventions in Social Science and Humanities https://doi.org/10.58806/ijiissh.2026.v3i1n12
Elmahjub E and others, “IslamicLegalBench: Evaluating LLMs Knowledge and Reasoning of Islamic Law across 1,200 Years of Islamic Pluralist Legal Traditions” [2026] Open MIND
Idrisov H and Ibragimov R, “Fatwa Issuance in the Islamic Law: The Specific of Implementation under the Conditions of the Chechen Republic (History and Modernity)” (2023) 172 SHS Web of Conferences http://dx.doi.org/10.1051/shsconf/202317204001
Usmonov M, “From Human Scholars to AI Fatwas: Media, Ethics, and the Limits of AI in Islamic Religious Communication” (2025) 5 Al-i’lam – Journal of Contemporary Islamic Communication and Media https://doi.org/10.33102/jcicom.vol5no1.125
Xu Y and others, ‘Artificial Intelligence: A Powerful Paradigm for Scientific Research’ (2021) 2(4) The Innovation 100179 https://doi.org/10.1016/j.xinn.2021.100179
Reports and Policy Papers
European Parliament, Artificial Intelligence: Potential Benefits and Ethical Considerations (Briefing, PE 571.380, 12 October 2016)
