Indonesia Weighs Risk-Based Model for AI Regulation

Indonesia’s current approach to artificial intelligence regulation is under review, with academic and policy discussions focusing on risk-based frameworks as a potential path forward.

At the 4th International Conference on Law, Governance, and Social Justice in 2026, Dr Parulian Paidi Aritonang, Dean of the Faculty of Law at the University of Indonesia, advocated for Indonesia to consider a risk-based model for AI oversight. He referenced the European Union’s Artificial Intelligence Act as an example but stated that Indonesia should adapt, not replicate, its provisions.

Dr Parulian described a regulatory approach that distinguishes AI uses by risk level, from prohibited and high-risk systems to those requiring transparency or minimal oversight. He outlined that high-risk AI systems would need to meet requirements in areas such as risk management, data governance, technical documentation, human oversight, and cybersecurity.

He also noted that the EU’s AI Act includes transparency rules for human-AI interaction, synthetic content, and deepfakes, as well as obligations for general-purpose AI systems covering documentation, copyright, and training data summaries.

What the numbers show

• 2026: Multiple academic studies published on Indonesia’s AI governance.
• March 2026: BI Institute working paper recommended a hybrid AI regulatory model.
• June 2026: Journal of Indonesian Legal Studies proposed risk-based liability for AI.

Indonesia currently relies on sectoral regulations related to electronic systems and personal data, but does not have a comprehensive law specifically for artificial intelligence. Academic publications in 2026 identified gaps in the country’s AI governance, citing outdated or fragmented legal instruments.

Several studies published in 2026 recommended that Indonesia move toward risk-based frameworks, independent oversight, transparency, and algorithmic auditing. These studies also suggested aligning Indonesia’s approach with international models, such as the EU AI Act, while adapting them to local needs.

A June 2026 article in the Journal of Indonesian Legal Studies called for changes to Indonesia’s AI liability framework, including strict liability, shifting the burden of proof, and mandatory transparency. Another study in Buletin Adalah concluded that the absence of a specific AI law increases legal uncertainty and risks, and called for risk-based regulation and independent oversight.

In the financial sector, a March 2026 working paper from the BI Institute recommended starting with an adaptive regulatory sandbox, transitioning to a risk-based framework, and eventually adopting a co-regulatory model that emphasizes human accountability through a Clear Box Liability framework.

Additional academic proposals include strengthening regulatory authority, implementing compliance auditing and certification, and ensuring institutional accountability, drawing from international principles and ethics recommendations.

* This article is based on publicly available information at the time of writing.