Để trí tuệ nhân tạo vận hành an toàn | VTV24

Key Concepts:

• AI Governance Framework: Establishing a robust legal framework for Artificial Intelligence (AI) management.
• Legal Basis for AI Data: Ensuring AI training data is managed by competent state authorities and reflects Vietnamese law, ethics, and culture.
• Ethical and Cultural Training for AI: Imbuing AI with an understanding of Vietnamese legal standards, ethical norms, and traditional cultural values.
• AI Risk Mitigation: Implementing mechanisms to support organizations and individuals testing AI, including insurance.
• Preventing AI Misuse in Public Policy: Prohibiting the abuse of AI to manipulate public policy for private interests.
• Specific Regulations for Sensitive Sectors: Developing tailored standards for AI in areas like healthcare.
• AI in Healthcare: Addressing the need for specific AI regulations for medical devices and diagnostic/treatment tools.

AI Governance and Legal Framework

The discussion emphasizes the need for a stringent legal framework for AI governance, aligning with Vietnam's practical realities. A key proposal is to establish a legal principle that the AI database, managed by authorized state agencies, must comprehensively contain information to train AI on Vietnamese law, ethical standards, and positive traditional cultural values. This is described as a "legal and technical golden hoop" to prevent AI from deviating from intended paths.

Mitigating AI Risks and Supporting Innovation

To address potential risks associated with AI, there's a call for state support mechanisms for organizations and individuals involved in AI testing. This includes the suggestion of an insurance mechanism prior to testing. Without such support, businesses developing high-risk AI systems, such as self-driving cars, especially in their startup phase, would be hesitant to conduct necessary trials.

Preventing Misuse of AI in Public Policy

A significant concern raised is the potential for AI misuse to "steer" public policy, serving the interests of developers or owners of the technology, who are often located abroad. To counter this, a proposal is made to include a specific article in the draft law prohibiting the abuse of AI in the public sector.

Specialized Regulations for Sensitive Sectors: Healthcare

It is argued that sensitive sectors like healthcare require supplementary, specific standards and regulations. This is to ensure that technological development proceeds in parallel with the safety and rights of citizens. Currently, AI-based medical products for diagnosis and treatment follow general medical device licensing procedures but lack specific AI standards and regulations. The hope is that the draft AI law will incorporate more concrete legal provisions for managing AI products in healthcare.

Synthesis and Conclusion

The core takeaway from the transcript is the urgent need for a comprehensive and nuanced legal framework to govern AI in Vietnam. This framework should not only address technical aspects but also deeply embed Vietnamese legal, ethical, and cultural values into AI development and deployment. Key priorities include ensuring AI data integrity, providing support for AI testing to foster innovation while managing risks, and strictly prohibiting the misuse of AI in public policy. Furthermore, the transcript highlights the critical necessity of developing sector-specific regulations, particularly for sensitive areas like healthcare, to guarantee the safe and beneficial integration of AI for the public.