The White Paper published by the UK Government’s Department for Science, Innovation and Technology (DSIT) on 29 March 2023 sets out its proposed approach to the regulation of AI in the UK. The White Paper, "Policy paper: a pro-innovation approach to AI regulation", follows on from the UK Government’s previous consultation undertaken in 2022.
The proposed framework is intended to serve as a guide to regulate AI technology, providing principles to increase public trust in AI developing capabilities. Instead of introducing new legislation (in contrast to the approach taken by the European Union), the rationale of the White Paper is to support regulators (such as the Information Commissioner’s Office, the Competition and Markets Authority, Health and Safety Executive, the Equality and Human Rights Commission and the Financial Conduct Authority) in overseeing the development of AI within their respective areas of responsibility.
The DSIT states that they will keep the proposed approach under review and will evaluate whether the application of the principles is effective and consistent among regulators. This is intended to address earlier concerns over likely regulatory inconsistency, particularly in relation to organisations that work across more than one sector, and which may need to be compliant with more than one regulator.
The consultation runs for 12 weeks until 21 June 2023 and interested parties have an opportunity to provide feedback before then. Subsequent to the consultation, consideration will be given to whether legislation is necessary to facilitate compliance with the guidance.
There are five principles cited in the White Paper for regulators to consider when building trust and providing transparency in AI innovation. These principles will be translated into guidance issued by UK regulators over the course of 12 months, including risk assessment templates and other tools.
The principles are as follows:
- Safety, security and robustness
- Transparency and explainability
- Accountability and governance
- Contestability and redress
In addition, the following four elements of the AI framework are designed to provide a clear regulatory approach:
- Defining AI based on its unique characteristics to support regulator co-ordination.
- Implementing a context-specific approach.
- Providing cross-sectoral principles to assist regulator responses to AI risks and opportunities. Initially, regulators will be able to apply these principles at their own discretion to meet the needs of their respective sectors. Thereafter, the UK Government expects to impose a statutory duty on regulators to adhere to the principles when parliamentary time allows.
- Supplying new central government functions to support regulators in their implementation of the AI regulatory framework and the benefits of a regulatory approach.
Annex A of the White Paper summarises the factors regulators may wish to consider in developing their guidance.
The UK Government will publish an AI regulation roadmap which will have further information on the implementation of the regulatory framework and include details of the central government functions recognised as one of the four elements to the AI guidance. This roadmap will be published alongside the UK Government’s response to the consultation on the White Paper.
Looking further ahead, the UK Government will publish a draft central, cross-economy AI risk register for consultation, an updated AI regulation roadmap and an evaluation report. The purpose of the evaluation report is to establish whether statutory intervention is required to aid compliance with the AI framework.
Lastly, the White Paper announced that the UK Government will invest £2 million for a new sandbox to allow organisations to trial new AI products prior to market launch and test how AI regulation would apply to their products.
This article was co-written by Arina Yazdi, Trainee Solicitor.