UK Government Proposes New AI Regulatory Regime

prologue

The UK Government (the ‘Government’) has published a proposal for a new regulatory framework for artificial intelligence (the ‘White Paper’). The goal is to make the UK one of the “top places in the world to build foundational AI companies” by “providing a clear and innovative regulatory environment”. The whitepaper outlines a common-sense approach aimed at fostering innovation without compromising security or privacy.

Regulators and industry have expressed concern about the risks posed by artificial intelligence (AI). In particular, an open letter endorsed by hundreds of influential figures discouraged the development of AI systems that “compete with humans,” citing concerns about harmful disinformation, job automation, and human obsolescence. I’m letting Italy’s data protection regulator temporarily banned his ChatGPT over data protection concerns, and the UK’s Information Commissioner fined US firm Clearview for its AI-powered facial recognition software. was imposed.

The European Commission has proposed its own draft regulation (the “AI Regulation”) harmonizing AI regulation across the bloc. This broadly prohibits AI applications with an “unacceptable risk” level. This article reviews the white paper’s proposals and highlights their differences from AI rules.

Proposed British regime

The UK regulatory framework is based on five key principles:

• Safety, Security, Robustness – AI systems must function in a robust and secure manner throughout their lifecycle.
• Appropriate transparency and “explainability” – AI systems should be well explained and fully transparent.
• fairness – AI systems must not violate the legal rights of individuals or organizations, unfairly discriminate against individuals, or produce unfair market outcomes.
• Accountability and governance – AI systems must be effectively monitored and clear accountability established.
• Competitiveness and Redemption – Stakeholders should be able to challenge AI decisions that risk harmful consequences.

The key principles are initially non-statutory, and regulators are advised to refer to them only when determining AI regulation. However, regulators may be obliged to consider key principles after an initial implementation period subject to government review.

Comparison between UK and EU

The government has taken a “common-sense, results-oriented approach” that balances the goal of making the UK an AI superpower by 2030 with the goal of “proportionally regulating” the significant risks posed by AI. I am proposing. The UK takes a markedly different approach than the EU.

• broader definition. The white paper defines AI as a system with capabilities that combine “adaptive” and “autonomous.” A broad definition allows regulations to be quickly applied and amended to evolving technologies. Instead, AI rules use a fixed, narrower definition of AI.
• Context specific. The white paper avoids product-specific categorization and instead favors outcome-based regulation. AI applications are contextualized and considered high-risk or low-risk on a case-by-case basis. Regulators use this to balance benefits and risks. In contrast, the AI ​​Regulation proposes a formal taxonomy of risks (from low to high) and prohibits the use of AI techniques deemed to pose “unacceptable” risks.
• No new regulator. The decentralized UK regime will not create a new regulator. Instead, it relies on existing regulators to enforce the Financial Conduct Authority, Office of the Information Commissioner, Competition and Markets Authority, Equality and Human Rights Commission, and Pharmaceuticals and Healthcare Products Regulatory Agency. In contrast, the EU proposes both a central regulatory authority (the European Commission for Artificial Intelligence) and national supervisory authorities in each member state.
• No new penalties. The whitepaper does not propose new penalties for violations. Instead, various UK regulators will use their current enforcement powers. In contrast, the EU regime has introduced new penalties of up to 6% of annual global turnover or €30 million (whichever is higher).
• role of governmentThe government provides central support functions for UK regulators, including monitoring effectiveness and consistency, assessing economic impact and facilitating interoperability with foreign governments. The white paper suggests that the government work with regulators to implement its proposals. Its role includes identifying risks in AI applications, working with regulators to identify cross-sectoral “core risks”, and implementing legislation and regulations to respond to risks identified by regulators. Including updating regulations.

potential problem

This white paper raises several issues.

• Do existing regulators have the capacity to assume additional responsibilities?
• Do existing regulators have the expertise to review compliance measures for technically complex AI applications?
• Will the regulatory alphabet soup create inconsistencies in the application of key principles, duplications and omissions in regulation?

The UK is considered one of the most ‘AI-ready’ countries, and its proposed principle- and results-based approach, underpinned by behavioral economics, has been successful in other contexts such as financial services. increase. Governments have adopted similar approaches to digital regulation, particularly with respect to digital platforms, data privacy and online safety.¹The Digital Markets Division within the Competition and Markets Department will be formally launched in the coming months, supported by legislation.².

As both the UK and EU become more aggressive in AI regulation, The two different systems may not only create conflicts, but also create additional compliance burdens, especially for companies doing business in the EU.

next step

The government has initiated a public consultation on the proposal for the white paper, inviting stakeholders and the public to respond to consultation questions by the June 21, 2023 deadline.

Thanks to trainee Ben Hasson who also contributed to this article.

¹ Digital Regulation: Driving Growth and Unleashing Innovation – GOV.UK (www.gov.uk).

²The Digital Markets, Competition and Consumer Bill will undergo legislative review this year.