Artificial Intelligence | UK Regulatory Outlook January 2024

EU AI Act

On 8 December 2023, negotiators for the European Parliament and Council of the EU reached political agreement on the shape and contents of the ground-breaking EU regulation on AI – the AI Act. However, this represents the beginning of the end of the legislative process, rather than the end itself as much of the detail remains to be finalised, including areas that could still provoke disagreement.

The text that was agreed in December is not publicly available but our Insight summarises what we knew at the stage when the agreement was reached. Since then, the Commission has published this useful Questions and Answers document giving an overview of the Act. We understand that the final text must be finished by the end of February in order for there to be sufficient time for the formalities of adoption of the text by the Parliament and Council to take place before the European Parliament elections in early June. The full and final text will be published as part of the transparency around those votes.

We anticipate that the definitive text legislation will be published in the EU's Official Journal in around five to six months' time. It will become law 20 days after publication – likely to be late spring/summer 2024.

A staggered compliance period will follow after the AI Act becomes law:

• prohibitions on certain categories of AI will come into force after six months (the end of 2024);
• the provisions on general purpose AI will come into force after 12 months (summer 2025);
• most of the remainer (including the high risk AI regime) will come into force after 24 months (summer 2026); and
• the provisions working AI regulation into existing product safety regimes will come into force after 36 months (summer 2027).

AI liability directive

We do not expect the AI Liability Directive to be finalised before the EU elections taking place in June 2024, as there is not enough time. It will be for the new Commission (to be appointed by the newly elected Parliament) to decide whether to continue with this legislation – although it seems likely that it will. We expect the destiny and timings of this legislation to become clearer in the second half of the year.

Possible legislative initiative on AI in the workplace 

Over the last couple of years, the EU Commission and European Parliament have been looking at the impact of AI in the workplace. It was the focus of the Commission's European Employment & Social Rights Forum 2023 event in November 2023. The new Commission may decide to develop these discussions into a new legislative initiative.

UK AI white paper

The UK government consulted on its white paper on AI last year and we currently expect its response to the consultation to be published early in 2024 (having been delayed from the end of 2023). The response is expected to set out the current UK government's policies around AI. The core of its approach is not to develop new legislation (and none was included in the King's Speech last November) but to issue high-level principles to guide the deployment by existing regulators of their existing powers.

Once the government's overarching strategy and principles have been published, we can expect to see more developed policies from the various UK regulators on how they are going to address AI within their particular areas of jurisdiction.

UK elections: a transition year for policy making?

There will be a general election in the UK by the end of January 2025 at the latest, with many commentators expecting it to take place in October 2024. If the current governing party were to be replaced, the new government is most likely to be formed by the Labour Party. Accordingly, if the elections take place at the expected date, this will be a transition year for policy making.

Our recent Insight on AI regulatory developments includes discussion of what might be expected from a Labour government in relation to AI. The party has stated AI to be a strategic priority and has indicated that it would look to provide "greater regulatory certainty" – suggesting new legislation?

IP negotiations

As we discussed in this Insight, the UK government has been working on a code of conduct around the interaction between AI, on the one hand, and intellectual property (IP) rights on the other. Many AI models – including powerful generative AI models – are understood to have been trained on data scraped from the internet, some of which may be protected by IP rights. However, there is debate around whether existing English law IP rules permit this. Copyright and database right holders consider that it does not, but AI developers consider it essential to the development of this transformative technology that they are able to use publicly available information from the internet. The code currently under negotiation is intended to find a path through this issue.

We expect the negotiations on the code to conclude this year.

Next AI safety summits and the development of international initiatives

At the first AI Safety Summit held in November 2023 in the UK, it was announced that the next summits are planned to take place in South Korea in six months and in France in 12 months.

In terms of other international developments on AI, such as the G7's code of conduct for AI developers and guiding principles on AI (see our previous Regulatory Outlook and this Insight), we will start to understand the effect of these international initiatives – which mostly depend on voluntary adherence – and how much impact they will have.