Vietnam is preparing to become one of Asia's first nations with a dedicated AI law, following the release of a draft bill that mirrors key elements of the EU's AI Act. The proposal aims to consolidate rules for AI use, strengthen rights protections and promote innovation.

The law introduces a four-tier system for classifying risks, from banned applications such as manipulative facial recognition to low-risk uses subject to voluntary standards. High-risk systems, including those in healthcare or finance, would require registration, oversight and incident reporting to a national database.

Under the law, companies deploying powerful general-purpose AI models must meet strict transparency, safety and intellectual property standards. The law would create a National AI Commission and a National AI Development Fund to support local research, sandboxes and tax incentives for emerging businesses.

Violations involving unsafe AI systems could lead to revenue-based fines and suspensions. The phased rollout begins in January 2026, with full compliance for high-risk systems expected by mid-2027. The government of Vietnam says the initiative reflects its ambition to build a trustworthy AI ecosystem.