Taiwan Unveils Draft Bill on AI

In a pivotal move for the AI industry, Taiwan’s National Science and Technology Council (NSTC) has introduced the draft bill of the Basic Act on Artificial Intelligence (AI) (人工智慧基本法草案). Announced on July 15, 2024, this groundbreaking legislative framework is now open for public consultation over the next 60 days, inviting stakeholders to provide feedback and shape the future of AI governance in Taiwan.

Key Highlights and Core Principles

Regulatory Framework and Government Role

The draft bill outlines essential principles for regulating AI development and application, reflecting a comprehensive approach to governance. The government’s role is emphasized in fostering AI innovation, with a focus on sustainable development, human autonomy, and privacy. The bill also proposes a review of existing laws, potentially designating a new or existing governmental agency to oversee AI regulation and coordination.

Public-Private and International Collaboration

One of the bill’s key features is promoting public-private partnerships and international cooperation. By establishing a regulatory sandbox, Taiwan aims to create an environment conducive to AI research, development, and application. This initiative is poised to accelerate innovation while ensuring robust regulatory oversight.

AI Literacy and Workforce Readiness

Recognizing the transformative impact of AI on the labor market, the bill underscores the importance of AI literacy and capacity building. Educational initiatives will prepare the workforce for changes brought by AI advancements, ensuring a seamless transition and minimizing societal disruption.

Risk Classification and Global Standards

The bill mandates the development of an AI risk classification framework aligned with international standards. This global interface will enhance Taiwan’s ability to manage AI-related risks effectively, fostering a safer and more reliable AI ecosystem.

Defining AI

The Taiwan bill defines AI as “a machine-based system capable of operating autonomously and, through inputs or sensing, providing outputs such as predictions, content, recommendations, or decisions that influence real or virtual environments.” This definition aligns with contemporary trends in AI regulations from the European Union, the OECD, and the United States, ensuring consistency with global best practices.

Core Principles of AI Governance

1. Sustainable Development and Well-being: Inspired by the Hiroshima Process Code of Conduct, this principle emphasizes responsible AI management to benefit humanity and the planet, promoting social fairness and environmental sustainability.
2. Human Autonomy: Respecting human rights and cultural values, the bill advocates for human supervision and rule of law, referencing the OECD’s recommendations to support autonomy and fundamental rights.
3. Privacy and Data Governance: Drawing from the US Blueprint for an AI Bill of Rights, the bill prioritizes data minimization and protection, promoting openness and reuse of non-sensitive data while safeguarding personal information.
4. Security and Safety: Ensuring the security and safety of AI systems, the bill references the US Blueprint and Singapore’s Model AI Governance Framework, aiming to protect against AI-related threats.
5. Transparency and Explainability: Following the EU’s Ethics Guidelines for Trustworthy AI, the bill requires suitable disclosure and labeling of AI-generated results to improve trustworthiness and assess potential risks.
6. Fairness and Nondiscrimination: To eliminate bias and discrimination, the bill mandates inclusive AI development, considering societal diversity and avoiding discrimination against specific groups.
7. Accountability: Ensuring internal and external governance responsibilities, this principle is formulated with reference to Singapore’s AI Governance Framework.

Government Missions

The bill sets forth the government’s missions to promote AI under these core principles. Key actions include:

• Providing financial incentives and support to AI-related industries.
• Establishing a regulatory sandbox for AI development.
• Enhancing public-private and international cooperation.
• Ensuring data quality and protecting intellectual property rights.
• Developing AI risk classification frameworks in line with international standards.

Conclusion

This draft bill marks a significant stride in Taiwan’s efforts to regulate AI, laying a robust foundation for future laws and directives. Industry players should stay abreast of these developments, preparing to align their operations with the evolving regulatory landscape.

As Taiwan charts its course in AI governance, the world watches, anticipating the innovations and advancements that will emerge from this forward-thinking legislative framework.

Source: K&L gates