AI, data and analytics update – May 2026

This blog explores recent legal, policy and industry developments in AI, data and analytics.

UK government launches Sovereign AI Fund

The fund is backed with £500m to invest in British AI startups, aiming to help the UK become “an AI maker, not just an AI taker”. At the launch, Tech Secretary Liz Kendall said “Sovereign AI… will combine the speed of venture with the strength of the nation state... It is independent. With an investment committee making its own decisions free from political interference about who to back and why.”

UK aims to “set the standards” for how AI is deployed

In a speech to the Royal United Services Institute on 28 April, tech secretary Liz Kendall said about the government’s approach to AI sovereignty involved:

• “Backing more British AI companies.” (via the Sovereign AI Fund)
• “Working more closely with our international partners, particularly other so-called middle power nations, including on setting the standards for how AI is deployed.”

Kendall praised the work of the AI Security Institute and noted that “at the next meeting of the international network of AI Security Institutes, which the UK Chairs, [it] will publish best practice on the science of evaluating AI models.”  

Anthropic’s Mythos Preview model raises concern about cybersecurity arms race

Anthropic announced it had developed (but would not publicly release) a new AI model which is adept at finding flaws in software.

 The fear is that Mythos Preview (and AI models like it) could increase the prevalence of cyber-attacks by making it much easier to identify vulnerabilities to exploit.

Anthropic predict that “the fallout – for economies, public safety, and national security – could be severe.” They have launched Project Glasswing with the likes of AWS, Google and Microsoft in what they say is “an effort to secure the world’s most critical software.”

Building your AI skillset: Mills & Reeve’s Iain Murdoch, head of legal AI, shares his thoughts with The Law Society

Iain’s top tips:

• Take a people-first approach to building your team’s AI expertise.
• Teach what key terms mean (eg, AI, generative AI, machine learning).
• Set “golden rules” which explain which AI tools can be used at work and when.

An AI assistant that promises to answer UK citizens’ questions based on government website content

On 16 March, the Government Digital Service published a blog detailing the results of a comprehensive pilot for an AI assistant that answers user queries by drawing solely on guidance published on the GOV.UK website. We are told that the AI assistant is 90% accurate and will be rolled out to users of the GOV.UK app in 2026, with further tests planned before a wider release on the GOV.UK website. Planned use cases involve answering queries on tax, benefits and visas.

Government releases AI and copyright report: What are the commercial implications?

Does the law require AI companies to seek permission and pay for a licence to use copyright works to train their AI? Should it? This is the crux of the global debate over how to apply existing copyright laws and whether to reform them.

The UK government report, published on 18 March, confirms that it will continue to take a watch and wait approach, preferring to “take the time needed to get this right”. This means there will be legally and commercially challenging decisions ahead for both AI companies and rights holders.

Lord Holmes continues to bang the drum for AI regulation

Lord Holmes published a “Parliamentary One-Pager” making the case for a standalone AI Bill. The government’s current stance is that it prefers to amend existing legislation in a piecemeal way, dealing with issues as they arise.

If you would like to discuss any of the matters in this update, or need legal advice connected to the development or deployment of AI please get in touch with our team of sector specialists. 

Our content explained