Click on image to download
In July 2023 the UK Digital Regulation Cooperation Forum (DRCF) issued a request for comment on the status of generative AI and priority regulatory concerns. The DRCF is comprised of four of the UK’s major regulators: the Competition and Markets Authority (CMA), the Financial Conduct Authority (FCA), the Information Commissioner’s Office (ICO) and Ofcom. As a research centre with a focus on technology regulation and evidence-based policy making, CREATe is undertaking research that addresses the specific questions raised by the DRCF. As we lay out in our response, published today as a CREATe working paper 2023/08, the challenges posed by generative AI will be complex, imposing a need for regulators to actively engage with evidence on these dynamic and multifaceted effects.
A range of regulatory concerns has emerged since generative AI became more visible with the launch of ChatGPT in November 2022. They include (1) Safety: outputs may be plausible and apparently well-grounded but wrong, dangerous or fraudulent; (2) Bias: output reflects input (e.g. from Internet scrapes, such as Common Crawl); (3) Black boxes: sources of information and how models were trained and aligned is often unknown; (4) Personal data: implicated both with respect to input and output; (5) Intellectual Property: training data may include content that is copyright protected; (6) Impacts on inter-firm competition and industry structure.
Addressing these issues requires an interdisciplinary approach. Regulatory approaches will need to be cross-domain, adaptable, and scalable to confront these multi-faceted challenges, supported by independent research.
The text of the working paper can be downloaded here.