EU AI Act agreed as UK says it won't legislate until timing is right

After months of fierce debate, political agreement was finally reached on the EU’s landmark AI Act on 8^th December 2023. Ursula von der Leyen described it as “the first-ever comprehensive legal framework on Artificial Intelligence worldwide” and it is true to say that it has caught the attention of organisations across the globe, with its wide extra-territorial reach, high fines and prohibition on certain types of AI which pose an unacceptable risk. 

It had looked like issues such as real-time biometrics and rules on generative AI could scupper the deal, and final details are still being ironed out at a technical level, but a very long last Trilogue session got it over the political line.

The Act (which will be a Regulation with direct effect in member states) will lay down a uniform legal framework for the development, marketing and use of artificial intelligence in line with EU values. The Commission confirms that its aim is to foster responsible innovation in Europe - “[by] guaranteeing the safety and fundamental rights of people and businesses it will support the development, deployment and take-up of trustworthy AI in Europe.” 

As well as laying down rules around the development and use of AI, the Act:

• Regulates according to risk – there are 4 categories: unacceptable risk which is prohibited; high risk which is heavily regulated; limited risk which is subject to transparency obligations; and minimal risk which has no additional obligations. 
• Has a wide scope - both in terms of extra-territorial risk and the fact that it covers a number of players in the AI supply chain.
• Contains innovation friendly provisions, such as sandboxes. 
• Sets high fine limits – a maximum of of €35 million or 7% of global annual turnover in the previous financial year.

More details on the Act, and its progress can be found in our client briefing, and my previous blog.

While the EU is pressing ahead with legislation, the UK has confirmed that it will not legislate on AI until the timing is right. On 13^th December, Michelle Donelan was asked by the Science Innovation and Technology Committee whether the UK would place a statutory duty on regulators to have due regard to 5 AI Principles set out in the Government’s AI White Paper (see our blog). Her answer (reported by Computer Weekly) was that timing was key with legislation to avoid stifling innovation, that the UK Government was taking action now and working hard to understand the risks, and that regulators already had legislation to follow. She went on to say that the UK was taking an evidence based approach and would not “lurch to legislate”, adding “[w]e have seen the impact that that can have, look at the EU AI Act and the ramifications and the response by industry to that…This isn’t as simple as ‘Legislation is the only tool in the toolbox’ – it certainly isn’t, and there are downsides with legislation, [like] the fact that it takes so long.”

For more information on AI generally, see our Regulating AI Series.