Vietnam’s AI Law: What It Means for Mental-Health Technology in 2026

 A new chapter for responsible AI

Illustration created with Canva AI by Author — symbolising Vietnam’s push toward ethical, human-centred AI under its 2026 digital technology law.

Vietnam is entering a historic moment in digital policy. The Draft Law on Artificial Intelligence, released for public consultation in late 2025, marks the country’s first attempt to build a comprehensive legal framework for AI. Once enacted in January 2026, it will guide how AI is developed, deployed, and supervised across industries — from fintech to healthcare.

For mental-health professionals and app developers, this law isn’t just about compliance. It’s about trust — ensuring that the algorithms supporting therapy, diagnosis, or emotional-wellbeing tools truly serve people, not the other way around.

---

 What the law says — in plain language

At its heart, Vietnam’s AI law is about balance: innovation with protection.
It introduces a risk-based framework inspired by the EU’s model, dividing AI systems into four categories:

Risk level	Examples	Regulatory focus
Unacceptable	Manipulative AI, invasive surveillance	Banned outright
High risk	Healthcare, education, finance	Registration, audits, human oversight
Medium risk	Chatbots, recommendation systems	Transparency & labeling
Low risk	Entertainment or creative tools	Voluntary codes of conduct

For mental-health technology, most products will fall under high-risk or medium-risk categories, depending on their purpose.

• A diagnostic AI that predicts depression risk → high-risk.

• A journaling or mindfulness chatbot → medium-risk.

This classification decides how much human oversight, documentation, and transparency each product must include.

---

Why mental-health AI needs special care

Mental-health tools touch some of the most private corners of a person’s life. The new law recognises this by enforcing human-centric principles: fairness, safety, transparency, and explainability.

That means:

• Users must know when they are talking to an AI.

• AI systems should never replace human clinicians in decision-making.

• Data used for mental-health prediction or therapy must respect privacy laws and consent boundaries.

In simple terms, the government wants to protect users’ rights while encouraging digital care innovation — a balance essential to public trust in AI therapy and wellbeing apps.

---

Human oversight and “AI with empathy”

One of the most discussed points is the requirement for human oversight. High-risk AI systems — such as those analysing emotions, recommending treatments, or detecting suicidal ideation — must always have a human-in-the-loop.

This is critical in mental health, where a wrong suggestion or delayed intervention could have life-changing consequences. The law encourages developers to design systems that assist rather than replace therapists — combining computational insight with human empathy.

---

Space for innovation

Importantly, the law doesn’t shut the door on creativity. It establishes AI sandboxes where start-ups can test new models safely, and offers tax incentives and funding through the forthcoming National AI Development Fund.

For local mental-health innovators — think mindfulness apps in Vietnamese, or rural tele-psychiatry tools — this is a major opportunity to build responsible, locally grounded technology.

---

Preparing for compliance

If you design or use mental-health AI systems in Vietnam, here’s how to get ready before the law takes effect in 2026:

1. Map your system. Identify what kind of AI you’re using and its potential health impact.

2. Assess risk. Classify it using the four-tier system.

3. Document everything. Prepare transparency reports, testing logs, and user-facing disclaimers.

4. Add human oversight. Always ensure a licensed practitioner can review critical outcomes.

5. Secure data. Protect health data under Vietnam’s cybersecurity and privacy laws.

6. Watch for decrees. Implementation guidelines will follow throughout 2026.

---

 Looking ahead

Vietnam’s AI legislation reflects a broader global trend: using regulation not to stifle progress but to guide it ethically. For the mental-health field, this offers a roadmap toward AI that heals, not harms — combining compassion with computation.

As mental-health technologies become more embedded in everyday life — from mood-tracking wearables to cognitive-behavioural therapy chatbots — Vietnam’s careful approach could help ensure that digital care remains human at its core.

Source: Ministry of Science and Technology, Vietnam. Draft Law on Artificial Intelligence (2025), released for public consultation.
Read the full draft on DFDL’s website

Related reading: Algorithmikē Psychē (quarterly AI policy briefs) / The UK's Google Ruling: What Fairer Search Means for AI, Trust, and Mental Health.