On July 18, 2022 - the UK government released a policy paper proposing to establish a pro-innovation framework to regulate AI.
What is the AI regulatory approach proposed by UK ?
What are the major differences with the EU AI Act ?
The UK context : "AI superpower"
Uk has the ambition to be a “AI superpower” that leads the world in innovation and development.
In 2021, the UK was 1st in Europe and 3d in the world for private investment in AI companies ($4.65 billion) and newly funded AI companies (49).
And given the ambition, the trend to build this strategic asset will not stop in the near future: UK has the target of total R&D investment in AI reaching 2.4% of GDP by 2027.
Scope of the approach proposed to regulate AI
The definition of scope is slightly different from Europe.
Unlike the European AI Act, the UK government does not clearly define the boundaries of what constitutes AI: the UK government allows the relevant authorities to decide which technologies and systems fall within the scope.
As a result, it is less prescriptive than the EU on a strict definition of what constitutes an AI system, especially one classified as high risk.
Instead, the UK government proposes to include "fundamental characteristics of AI" to assess the scope of the AI regulatory framework and allow regulators to establish and evolve more detailed definitions at the level of their application.
Fundamental characteristics of AI for inclusion
According to the UK government, 2 major criteria are to be considered for inclusion in the scope
"the adaptativeness of the technology": assessment of the intent or explanatory logic that has not been explicitly programmed with human intent, but has been "learned" based on a variety of techniques (e.g., data training)
"the autonomy" of the system: ability to strategize, respond, and make decisions without the express intent or ongoing control of a human.
Innovation framework based on principles
The UK government proposes to establish an innovation-friendly framework for regulating AI based on key principles
Specific context. Regulation is based on the use of AI and the impact it has on individuals, groups, and businesses in a particular context, and delegates responsibility to regulators to design and implement proportionate regulatory responses.
Pro-innovation and risk-based. Focuses on solving problems where there is clear evidence of high-risk concern, real risk, or missed opportunities.
This approach is quite flexible and does not rely on a single framework with a fixed, central list of risks or formal risk assessment.
While Europe focuses also on risk based approach and high-risk systems, the UK offers a proportionate approach to support growth and innovation and avoid unnecessary barriers to business.
Proportionate and adaptable. They propose to develop a set of cross-sectoral or domain-specific principles tailored to the distinct characteristics of each technology, based on the OECD principles to be interpreted and implemented in practice by regulators.
Consistency. They will ensure that regulators can coordinate effectively to ensure consistency among their respective approaches to the principles, including, where possible, collaborating to interpret or implement the principles on a common or cross-sectoral basis.
This policy sets out the general principles for regulating AI with flexibility for application and interpretation left to business and regulators.
But it's only the beginning of the regulatory journey.
As next steps, the UK government may examine potential gaps in the application of the framework, will foster coordination between the different regulators, and will ensure that the framework is aligned with the UK vision for AI.
For more info: https://www.gov.uk/government/publications/establishing-a-pro-innovation-approach-to-regulating-ai/establishing-a-pro-innovation-approach-to-regulating-ai-policy-statement