AI Can Help, But Humans Must Not Stop Thinking

UNIVERSITI SAINS ISLAM MALAYSIA (USIM), 23 SEPTEMBER 2026 —

Universiti Sains Islam Malaysia (USIM) has flagged growing risks in Islamic finance from over-reliance on artificial intelligence, warning that automation could erode critical human judgment in Shariah compliance without proper safeguards.

The concern emerged during a panel discussion at the Scientists Tell Stories forum, where Prof. Emeritus Dato’ Dr. Musa Ahmad—former dean of USIM’s School of Syariah and Law—argued that while AI could streamline processes like sukuk structuring or takaful underwriting, it risked replacing "human conscience and experience" in fatwa interpretation.

"AI can process millions of financial transactions in seconds, but it cannot tell you whether a murabahah agreement is exploitative unless a human teaches it the moral weight of fairness," he said.

The debate follows a surge in AI adoption across Malaysia’s RM1.2 trillion Islamic finance sector, where institutions are using machine learning to automate Shariah screening of investments, assess sukuk risk profiles, and even generate draft fatwas for routine transactions. Bank Negara Malaysia’s 2023 Islamic Financial Services Industry report noted a 40% increase in AI-driven compliance tools among licensed Islamic banks since 2022, with Maybank Islamic and CIMB Islamic leading early implementations.

Yet the USIM panel highlighted blind spots. AI models trained on historical sukuk data may overlook evolving Shariah rulings—for example, the Accounting and Auditing Organisation for Islamic Financial Institutions’ (AAOIFI) 2021 guidance on digital asset compliance—or misclassify hybrid financial instruments. "A computer cannot weigh the maqasid al-shariah [higher objectives of Islamic law] when a wakalah agreement’s terms shift from agency to partnership," said Dr. Nor Azlin Mohamad Ali, a USIM Shariah economist.

She pointed to a 2022 case where an AI system at a Malaysian Islamic bank misflagged a mudarabah-structured real estate fund as non-compliant after interpreting rental yield projections as riba-adjacent, requiring a manual fatwa override.

The risks extend to takaful, where AI’s predictive analytics for claims processing may inadvertently bias underwriting against certain demographics—a violation of Islamic principles of equity. A 2023 study by the International Islamic Liquidity Management Corporation (IILM) found that 68% of Malaysian takaful operators using AI tools lacked human audits to verify algorithmic fairness. "The danger isn’t AI itself, but the illusion that it removes human accountability," said Dr. Musa.

Industry players acknowledge the tension. Maybank Islamic’s head of digital innovation, Datuk Seri Azman Ismail, told Reuters that the bank’s AI-assisted Shariah compliance suite—deployed in 12 sukuk issuances last year—includes a "human-in-the-loop" review layer. "We use AI to flag anomalies, but the final ijtihad [juristic reasoning] is made by our Shariah board," he said. However, smaller Islamic banks with limited Shariah committees may lack resources to challenge AI-driven decisions, raising systemic risks.

The panel’s warnings align with broader global trends. The Islamic Financial Services Board (IFSB) issued a 2023 advisory urging member states to establish "AI ethics boards" for Islamic finance, with Malaysia among the first to respond. Bank Negara’s Guidelines on Technology Risk Management (2024) now require licensed Islamic institutions to document human oversight in AI-driven Shariah processes—a step Dr. Musa called "a necessary but insufficient fix."

Looking ahead, USIM’s researchers propose integrating "narrative intelligence" into AI systems—teaching models to recognize not just financial data but the human stories behind transactions. For example, an AI analyzing a qard al-hasan loan could be programmed to flag cases where borrower hardship might warrant waived repayment terms, mirroring the discretion Shariah scholars apply. "We’re not asking for a return to paper ledgers," said Dr. Musa.

"But we must ensure AI amplifies human judgment, not replaces it."

The challenge mirrors global debates in conventional finance, where regulators like the Monetary Authority of Singapore have mandated "explainable AI" in banking. In Malaysia, the stakes are higher: a 2023 survey by the Global Islamic Finance Forum found that 72% of Malaysian Muslims distrust AI-driven financial advice, citing concerns over Shariah authenticity.

As AI tools proliferate, the sector’s ability to maintain public trust may hinge on whether institutions can balance efficiency with the irreplaceable role of human ijtihad.

Related: Prof. Emeritus Dato’ Dr. Musa Ahmad · Universiti Sains Islam Malaysia (USIM)