Released on 18 July 2022, the Policy Paper describes the UK Government’s proposed framework for regulating the use of Artificial Intelligence (AI).
The UK is proposing a pro-innovation framework for regulating AI that would be (i) context-specific (use and impact on individuals, groups and businesses), (ii) pro-innovation and risk-based (with a focus on high risk concerns rather than hypothetical or low risks), (iii) coherent (cross-sectoral approach to AI that will foster regulatory coordination) and, proportionate and adaptable.
With regard to the cross-sectoral approach to AI, the UK outlines 6 early proposals:
Ensure that AI is used safely: It will be important for all regulators to take a context-based approach in assessing AI-related risks and take a proportionate approach to managing these risks.
Ensure that AI is technically secure and functions as designed: The functioning, resilience and security of a system should be tested and proven, and the data used in training and in deployment should be relevant, high quality, representative and contextualised. This will foster consumer and public confidence in the proper functioning of AI systems.
Make sure that AI is appropriately transparent and explainable: Explainability remains an important research and development challenge. But explainability must be required in certain circumstances (i.e. to explain certain decision-making to the public, consumers and businesses). The Paper outlines that in some high risk circumstances, regulators should entirely prohibit decisions that cannot be explained.
Embed considerations of fairness into AI: To ensure proportionate and pro-innovation regulation, regulators shall continue to define fairness by considering their sectors or domains.
Define legal persons' responsibility for AI governance: Accountability for the outcomes produced by AI and legal liability must always rest with an identified or identifiable legal person - whether corporate or natural.
Clarify routes to redress or contestability: The use of AI should not remove an individual’s or group’s ability to contest an outcome (for example when the relevant training data reproduces biases). Regulators should implement proportionate measures to ensure the contestability of the outcome of the use of AI.
Currently, AI is partially regulated through a patchwork of legal and regulatory requirements such as the UK data protection law. Although, and as outlined in the Paper, the UK Government does not intend to introduce legislation to regulate AI at this time; the principles will likely be implemented on a non-statutory basis supplemented by clear guidance from the government.
A White Paper and public consultation is scheduled for publication in late 2022.
The UK Government’s proposed light-touch approach to regulating AI contrasts with the EU's more prescriptive approach as laid down in the AI Act proposal published on 21 April 2021. For more information on the European Commission proposal on a legislative framework on Artificial Intelligence read our previous article here