AI, Adab, and the Future of Islamic Education

If the state allows artificial intelligence to become a teacher without rules, we are surrendering part of educational authority to a system that is not publicly elected, not fully transparent, and cannot be held morally accountable. In Islamic education, the risk is greater. What is at stake is not merely the accuracy of answers, but the way a new generation understands the authority of knowledge, truth, and responsibility.

AI is already in the learning space. The Stanford AI Index Report 2026, citing the Chegg Global Student Survey 2025, recorded that 95 percent of Indonesian university student respondents use generative AI to support their studies, the highest among the 15 countries surveyed. BPS, in its 2025 Indonesian Youth Statistics, recorded that 96.69 percent of young people aged 16–30 have used the internet. Yet the device gap remains sharp: computer use among urban youth reached 31.57 percent, compared with 15.08 percent in rural areas; among youth with disabilities it was only 7.20 percent.

This means the AI revolution is arriving when digital access is not yet equal. The state must not only ask how quickly schools and madrasahs are adopting AI. The more democratic question is: who benefits, who is left behind, whose data is used, and who is responsible when the system errs.

The Ministry of Religious Affairs has begun to move. In August 2026, the ministry published 39 Islamic Religious Education and Character Education books as well as one Islamic Religious Education book for general higher education institutions, integrated with an AI-based Smart PAI. Its assistant is limited to answering based on validated official materials. Religious Affairs Minister Nasaruddin Umar reminded, “Do not let religious education fall behind the way they learn.” This step deserves appreciation because it shows that Islamic education does not reject technology.

But innovation is not enough when governance lags behind. Precisely when AI enters the domain of religious education, accountability standards must be higher. A system can produce answers that sound convincing without bearing the consequences when it is wrong. It can process text quickly, but it does not automatically understand social context, diversity of opinion, the impact of decisions, or ethical responsibility towards the human beings who receive its answers.

This is where adab must be elevated. Syed Muhammad Naquib al-Attas formulated education as the inculcation of adab in the human being. In the context of AI, adab must not stop at advice for students to “use technology wisely”. Adab must be translated into governance principles: sources must be traceable, important decisions must remain with humans, student data must be protected, bias must be audited, and those who develop systems must be held accountable.

This perspective aligns with the warning from UPI Professor Tatang Herman that students should not hand over academic tasks entirely to AI. Deputy Minister of Communication and Digital Affairs Nezar Patria also stressed that “technology develops exponentially. Ethics must not continue to move linearly.” The government is now preparing a Presidential Regulation on the National AI Roadmap and AI Ethics with principles centred on humans, transparency, auditability, and risk-based governance.

This momentum must not pass without a specific education agenda.

First, the Ministry of Religious Affairs, the Ministry of Primary and Secondary Education, the Ministry of Higher Education, Science and Technology, and the Ministry of Communication and Digital Affairs need to establish national AI standards for education. Platforms used by schools, madrasahs, pesantren, and universities must explain their knowledge sources, data use, system limitations, and correction mechanisms. For high-risk religious questions, AI must function as a tool, not a final decision-maker. The human-in-the-loop principle must be an obligation, not an option.

Second, learning and assessment design must change. If an assignment can be completed with a single prompt, what has failed is not only the student who copies the answer, but also an educational design that still measures old capabilities. Oral examinations, argument defence, source tracing, process portfolios, and case studies need to be expanded. Teacher training must also go beyond prompt engineering towards the ability to read bias, hallucination, privacy, copyright, and source validity.

Third, the state must build digital Islamic knowledge infrastructure as a public good. The works of Nusantara ulama, research from state Islamic higher education institutions, validated Islamic religious education sources, fatwas of official institutions, and pesantren heritage need to be digitised systematically, openly, traceably, and made usable for national AI development. Without this investment, Indonesia risks becoming a large market for machines built with the data, language, and priorities of other parties.

This is also a matter of democracy. UGM academic Pratiwi Utami reminded that those who control data and knowledge hold the power to determine how reality is produced. If education surrenders knowledge to closed algorithms without the public’s right to know how they work, we are creating a new authority without adequate accountability mechanisms.

The DPR and the government must therefore not wait for the first major case. The National AI Roadmap must include education as a high-risk sector, with protections for children, students, and religious data. The Ministry of Religious Affairs must also ensure that digital transformation does not merely produce platforms, but a national knowledge ecosystem that can be audited by the public.

The future of Islamic education is not determined by how many chatbots are present in the classroom. The measure is whether technology makes human beings increasingly…