The guard has changed in Britain. For the first time in 14 years, the Labour Party is back in power, led by Prime Minister Keir Starmer, who took office on July 5.
Starmer was set to introduce a long-awaited artificial intelligence bill last Wednesday as part of the King’s Speech, in which Charles III read out the new government’s agenda. But the AI bill was pulled at last minute from the address for undisclosed reasons.
We’ll take a look at Labour’s agenda for potential AI legislation — what they’re planning, when it could come, and how their focus will differ from their Tory predecessors. But first, let’s examine Rishi Sunak’s legacy and whether he accomplished his goal of being a global leader on AI.
What did Sunak accomplish?
Sunak’s crowning moment was the AI Safety Summit in Bletchley Park on Nov. 1-2, 2023. The summit, held at the famed World War II codebreaking facility, was a global gathering on artificial intelligence safety aimed at international cooperation to deter AI’s worst-case scenarios from occurring. The Bletchley Declaration, the resulting document, was signed by the UK, the United States, and the European Union, but also, notably, China, along with two dozen other signatories. (And Sunak got to pal around with tech CEOs such as OpenAI’s Sam Altman, xAI’s Elon Musk, and DeepMind cofounder Mustafa Suleyman — who has since joined Microsoft.)
Bletchley was an important international agreement, but it also signaled that the UK’s leadership — under Sunak, at least — would be light-touch. He didn’t call for AI legislation, preferring to deploy Bletchley as a voluntary global corporate and government agreement.
Nick Reiners, a senior geotechnology analyst at Eurasia Group, said the hyperfocus on “existential risk” of AI is somewhat of a niche Silicon Valley obsession, a crowd that Sunak was interested in appeasing. “He saw AI as a way to build a legacy in a short time and this issue was something that animated him personally.”
What’s on deck for Labour?
Scott Bade, also a senior geotechnology analyst at Eurasia Group, doubts Starmer will follow Sunak’s lead on AI, but said he won’t throw away the standing that Sunak won for the country either. “Starmer does not have a signature global issue yet, and is unlikely to see AI as that issue,” Bade said. “But I'd be surprised if the UK didn’t keep showing up at the table to build on what Sunak did since this is the niche Britain now has in AI global governance. It will just be dialed down a peg or two.”
Compared with Sunak’s existentialist concerns, Starmer should be more focused on the short-term harms of artificial intelligence, Reiners said, citing workers’ rights and bias as examples. And with that comes the promise of actual legislation.
The bill that Labour was set to introduce would have reined in the most powerful large language models — but actual regulation seems to have been pushed off. In his speech, King Charles read off bills about cybersecurity as well as digital information, which seem to have won out over the AI bill, at least for now. “My suspicion is that they opted not to present this [AI regulation] now as they didn’t want to upset their growth narrative,” one tech leader told the Financial Times. Reiners said that departmental limits on parliamentary bills per session could be a constraint as well, and that AI was a lower priority.
When an AI bill is introduced, expect it to still be light-touch relative to the more expansive European AI Act. “I would say the UK is still generally respected as taking a thoughtful innovation-friendly approach to regulation in general,” Reiners noted.
The UK is home to successful AI startups such as Stability AI, maker of the image model Stable Diffusion, Google’s DeepMind lab, and the digital avatar company Synthesia, which we profiled in last week’s edition. And big AI-focused US tech companies, such as Microsoft and Salesforce have recently invested in the country. With the country’s economy on the ropes, Starmer’s challenge is to introduce legislative reforms that won’t totally scare off Big Tech.