Vietnam’s new IP law clarifies data use for AI training and digital enforcement

Vietnam’s IP overhaul is a good example of second‑wave AI regulation: not a headline AI Act, but sectoral law quietly rewritten to accommodate machine learning realities. By clarifying how computer programs are protected and when data can be lawfully used for AI training and research, Hanoi is signaling that it wants to encourage local AI development without abandoning rights holders. Equally important, the law expands intermediary liability and safe‑harbor rules for online platforms, nudging them toward more proactive IP enforcement in digital environments.

For the AI industry, the details matter. Uncertainty around whether scraped or licensed datasets violate copyright is a major friction point for both startups and multinationals. Clearer statutory language about permissible training uses—paired with shorter publication and opposition timelines for IP filings—could make Vietnam a more attractive base for AI companies serving Southeast Asia. At the same time, tougher obligations on platforms mean that AI model providers and hosting services will face higher compliance burdens when dealing with user‑generated content or code.

In the race to AGI, moves like this don’t change the global timeline, but they do shape who gets to participate. Jurisdictions that modernize IP and data rules in a balanced way will be better positioned to host regional model labs, dataset consortia and application developers, rather than merely importing AI systems built under foreign legal assumptions.