Govt says AI chatbots aimed at children will have safety built in

Digital Minister Gobind Singh Deo announced that the government plans to introduce the Artificial Intelligence (AI) Governance Bill in the Dewan Rakyat early next year, rather than this year. The proposed legislation aims to mandate safety features for AI chatbots and apps targeted at children. Deo stated that child protection would be comprehensively integrated into the bill, which is being drafted as a foundational framework capable of adapting to evolving technology risks.

The bill would adopt a risk-based approach, mandating developers of children-oriented apps or chatbots to implement safety-by-design measures such as regular risk impact assessments and content filtering to prevent psychological manipulation of minors. The disclosure requirement, notifying users that they are interacting with an AI system instead of a human, would be enforced through a targeted mechanism rather than a blanket requirement.

The necessity for disclosure would be determined based on factors like the application’s risk level, sector-specific regulations, and sectoral codes of ethics to be issued by the regulator and the minister. Gobind highlighted that AI service providers would eventually need to display clear notices to inform users they are interacting with AI, particularly for high-risk applications.

The bill's legal controls would be reinforced by the National Guidelines on AI Governance and Ethics, the enforcement of the Online Safety Act, and digital literacy initiatives like the free CyberSAFE untuk Rakyat program. Deo's comments came in response to a question from Riduan Rubin regarding the safeguards in the AI Governance Bill to mitigate risks for children using AI chatbots and whether bot disclosure would be mandatory to protect minors' cognitive awareness.

The government anticipates tabling the AI Governance Bill in the Dewan Rakyat early next year, pending further drafting and assessment of the bill's level of control and obligations based on potential harm posed by AI systems.