How can the Government ensure their use of artificial intelligence (AI) operates lawfully, fairly, and without discrimination? Find out in our evidence to the House of Commons Science and Technology Committee’s inquiry into the governance of AI.

Read the full evidence submission

Securing meaningful transparency should be the first port of call when considering the regulation of AI: without transparency, there can be no evaluation of whether systems are working reliably, efficiently, and lawfully, including assessment of whether or not they unlawfully discriminate.

Current governance of AI in the UK does not provide an adequate level of transparency to ensure these crucial safeguards are in place.

While we recognise the potential of the ‘Algorithmic Transparency Standard’ (ATS) to increase transparency of the public sector’s use of algorithms, this potential will remain limited if engagement with the ATS is not made compulsory and if improvements are not made to ensure people can understand the decision making processes they are subjected to.

And if the Government’s White Paper on AI is to reform the legal framework, the focus should be on fortifying existing safeguards and ensuring clarity and coherence between existing laws.

Our recommendations include:

  • Making the use of AI more transparent and explainable to the public
  • Improving transparency in the ATS and making it compulsory for public sector organisations
  • Introducing an independent regulator for the ATS, such as the Information Commissioners Office
  • Reforming the  legal framework that governs AI, with the focus on fortifying existing safeguards and ensuring clarity and coherence between existing laws
  • Ensuring there are quick and effective ways to enforce existing rights within the legal framework that governs AI
  • Ensuring AI regulations are based on the following principles: anti-discrimination, relfexivity within Government, respect for privacy and data rights, meaningful transparency, and accountability and avenues for redress

Read the full evidence submission