AI 'Ustadz' Cannot Serve as a Religious Reference
As technology evolves to assist human activities, the use of artificial intelligence (AI) to answer religious questions has become a phenomenon easily embraced by the younger generation, according to the Ministry of Religious Affairs (Kemenag).
Muchlis M. Hanafi, Director of Islamic Religious Information at the Directorate General of Islamic Community Guidance, stated that young people are a digital generation accustomed to obtaining information quickly and instantly through various digital platforms. However, he cautioned that the speed of acquiring information in the digital space often surpasses the depth of understanding. Therefore, the younger generation needs digital literacy to be able to sort out religious information that is correct and accountable.
He explained that Islamic knowledge is not only related to religious texts but also involves context, methodology, and wisdom in its application. These aspects cannot be fully replaced by technology. For religious matters requiring legal rulings or fatwas, the public must still refer to scholars and religious institutions that hold authority.
Ayu Purwarianti, an AI expert from the Bandung Institute of Technology (ITB), explained that the workings of AI Large Language Models (LLM) like ChatGPT are based on ‘next token prediction’, where each word is generated based on probability values. The AI produces answers as a sequence of words with high probability without understanding the meaning of the sequence. This mechanism means AI still has the potential to generate incorrect answers or responses. The primary party responsible for the use of information obtained from AI is the user, who must therefore be critical of AI-generated information.
AI is a digital platform that provides information to users based on data and information from the internet. However, not all information on the internet is correct. This means that AI cannot even be considered a trustworthy source of information, let alone a religious reference for issuing fatwas. Users are expected to validate the information obtained by referring to trusted sources, such as scholars or directly to syar’i arguments.
This is crucial because incorrect religious information can lead to erroneous understanding and, consequently, misguided practice. If one’s practice is wrong, it is feared that it may lead to sin. Moreover, replacing sincere scholars with digital platforms under state supervision, where algorithms are designed based on policy and security criteria, could potentially produce filtered and pre-formulated answers.
Fundamentally, Islamic law and fatwas are derived from the Qur’an, Sunnah, Ijma’, and Qiyas through the process of ijtihad. Ijtihad is a serious effort by religious experts to reach a legal ruling on cases not explicitly addressed in the Qur’an and Sunnah. However, performing ijtihad requires detailed qualifications set by ushul scholars, including extensive study, correct understanding of texts, sufficient knowledge of the Arabic language, and a broad comprehension of syar’i matters and their proofs.
Therefore, referring to Islamic law or religion and requesting legal fatwas must be directed to rational and pious scholars, not to artificial intelligence which is prone to error. This is in accordance with the word of God: ‘We did not send any messenger before you (Muhammad), but as men to whom We gave revelation. So, ask the people of knowledge if you do not know’ (Q.S. an-Nahl: 43). Scholars provide legal information or fatwas based on syar’i evidence and fear of God alone. Thus, a digital platform that lacks reason and consciousness can never replace the position of scholars in issuing fatwas or serving as a religious reference.