Communications Ministry to Take Risk-Based Approach in AI Ethics Rules

Jakarta (ANTARA) — The draft Presidential Regulation on Artificial Intelligence (AI) Ethics will use a risk-based approach to classify AI technologies.

Irma Handayani, head of the Communications and Digital Ministry’s team for regulations, the roadmap, ethics, governance and oversight of AI, said the approach divides AI into high-risk, low-risk and unacceptable-risk categories.

“We group them into three categories. In the European Union, we see four, while here we have three: low, high and unacceptable risk,” Irma said during a discussion at the Peruri Digital Summit 2026 in Jakarta on Wednesday.

AI with the potential to threaten security and safety will be classified as an unacceptable risk. For the low-risk category, Irma cited chatbots commonly used to provide public services, such as information.

She said different sectors would take different approaches to assessing AI risks, with healthcare and agriculture among the examples.

The draft regulation therefore requires each ministry and government agency to set its own risk standards. Organisations and companies using AI in particular sectors will have to comply with the sector-specific rules established for them.

Irma added that the draft regulation also sets out a framework for monitoring and evaluating AI operations. Oversight of AI use will be encouraged through self-assessment mechanisms for operators in each sector.

“Through this, they will measure their level of maturity. It is not actually mandatory, but it adds value for them. As they become more mature, public trust will also increase,” Irma said.

As an additional safeguard, the regulation will require ministries, government agencies and sector operators to provide incident-reporting mechanisms to address technical issues and harm experienced by AI users.