A focus on remote problems and a lack of binding commitments means that the summit was more style than substance.
The UK’s AI Summit has drawn to a close, to a mixed reception and, probably, to limited effect.
The highlight was the Bletchley Declaration on AI safety, which called for international cooperation on the risks of so-called Frontier AI, particularly in the areas of cybersecurity, biotechnology, and disinformation.
More than two dozen countries have signed up, including the UK, US, EU nations, India, and China, and have agreed to work together to support a network of scientific research on AI safety.
UK Prime Minister Rishi Sunak was keen to claim it as a British success.
“The UK is once again leading the world at the forefront of this new technological frontier by kickstarting this conversation, which will see us work together to make AI safe and realize all its benefits for generations to come,” he said.
However, the declaration is by no means as ground-breaking as Sunak would like to imply, nor is it anywhere near as comprehensive as it might be. For a start, it’s focused almost exclusively on so-called Frontier AI, and thus on the more existential threats potentially posed by the technology rather than on the immediate issues.
Legislative efforts
Meanwhile, the EU is already in the final stages of passing its AI Act, which sets out detailed regulations based on the different risk levels of various applications. Earlier this week, the G7 agreed on a set of guiding principles for AI and a voluntary code of conduct for AI developers.
In the US, President Biden this week signed an executive order ordering risk reporting from AI firms, described in a White House statement as ensuring that “America leads the way in seizing the promise and managing the risks of artificial intelligence.”
And, at the UK summit itself, US Secretary of Commerce Gina Raimondo announced a new US-led international AI safety institute within the Department of Commerce in what it’s hard not to see as a spot of one-upmanship.
“Light on detail”
The focus on Frontier AI is something of a boon for the big AI firms, who would no doubt vastly rather see the conversation centered around future theoretical risks than on issues of data privacy, security, fairness, job losses, and the like.
And the declaration has been described by Professor Anthony Cohn of the University of Leeds, and foundational models theme lead at the Alan Turing Institute, as “heavy on vision, but, unsurprisingly, light on detail.”
It’s far more urgent, he says, to address short-term risks such as biased training data and the resulting biased output than “possible, but still distant, risks relating to the conceivable creation of AGI, which the declaration is more focused on.”
And while further meetings are planned, the declaration is unclear on how a ban on dangerous AI could be enforced – signatories such as China may have very different views to those of the EU.
“The Bletchley communique is welcome and illustrates the importance of AI, but it falls short of binding arrangements that will control and shape the type of AI being developed at such great pace,” says Dr. Andrew Rogoyski, director of innovation and partnerships at the Surrey Institute for People-Centred AI.
“The UK and US AI Safety Institutes are a sensible step forward, but my fear is that by the time they are established and understand how to operate, the world of AI will have already moved on.”
While Sunak acknowledged that AI might need statutory regulation in the future, his ‘landmark’ agreement with tech companies that they should test their models before release remains voluntary – and doesn’t include the agreement of China.
Sunak’s PR campaign
It’s hard not to see the declaration – and the summit itself – as something of a PR exercise. The tone was set when it was originally announced that the summit would be held at Bletchley Park, the highly evocative home of British codebreakers during the Second World War. Wheeling out King Charles to address the summit was part and parcel of this image-boosting effort.
Sunak is desperate to please tech firms – which may be the reason why delegates to the summit belonged overwhelmingly to the industry and to the government, rather than including civil society groups and unions.
The final event of the two-day summit was an interview between Sunak and Elon Musk, in which – almost unbelievably – they discussed the Terminator movies and the importance of an ‘off-switch’ for AI. Sunak appeared adoring throughout.
The summit has (mostly) been a good PR exercise for Sunak and has no doubt cemented his relationship with AI firms. But with real legislation emerging around the world and technology changing fast, will it improve AI safety around the world? Probably not much.
