英国科学创新技术部+支持创新的人工智能监管方法2023-91页_1mb

A pro-innovation approach to AI regulation

Background

• The UK is ranked 3rd in the Global AI Index, with 3% of global AI startups.
• Countries are establishing rules for AI, necessitating UK leadership in global AI governance.

Core Principles

• Safety, security and robustness: AI systems should function securely and reliably.
• Appropriate transparency and explainability: Users should understand AI outputs and how they work.
• Fairness: AI should not discriminate and should be proportionate.
• Accountability and governance: Clear lines of responsibility for AI decisions.
• Contestability and redress: Users should be able to challenge or appeal outcomes.

Implementation

• Regulations are context-based and not system-specific.
• Regulators will implement the principles proportionally within their remits.
• A non-statutory framework will be initially applied before possible legislation.
• A duty to “have due regard” may be introduced if needed to strengthen implementation.

Central Functions

• Monitoring and evaluation: Gathering evidence to assess effectiveness.
• Cross-sectoral risk assessment: Addressing risks that go beyond single regulatory scope.
• Regulatory support: Assisting innovators to navigate complex rules.
• Horizon scanning: Tracking AI trends and global standards.
• International alignment: Ensuring interoperability with global AI approaches to foster trade and cooperation.

Conclusion

• UK’s approach prioritizes innovation while ensuring responsible AI use and public trust.
• The framework is adaptable and future-proof to accommodate AI development.