In less than two weeks, the world’s second artificial intelligence (AI) safety summit will take place in Seoul. South Korea and the UK are co-hosting the two-day affair that will give air time to global discussions about the progress of AI.
Since Britain hosted the first summit of this kind six months ago, little has happened. But this could be about to change.
On Friday, the House of Lords will hear the third reading of the AI bill and, if all goes well, it will progress to the House of Commons.
“There have been both interesting and positive noises from Number 10 but also a development and more positive signs coming out of Labour as well in terms of their intentions for legislating for AI,” said the man behind the bill, Lord Chris Holmes.
Holmes, who introduced the bill to Parliament in November last year, believes some elements or a version of his legislation may appear sooner than initially expected, perhaps as early as this summer.
This may come as a surprise to some, given that since the first safety summit, which Holmes described as a “very positive event”, very little activity has happened.
Aside from a recent announcement of a UK partnership with the US‘ AI safety institute, there has only been a report into how the government will prep regulators to control the technology.
In February, discussions over an AI intellectual property code collapsed suddenly and were then entirely abandoned.
Holmes previously told City A.M. that MPs were growing restless over the slow government approach to creating checks and balances over AI, despite describing it as a potential “existential threat”.
The Conservative government has been taking a ‘light touch’ approach out of fear it will stifle innovation and investment in the UK but it has always said there will come a time to legislate AI.
“What we’re sensing is that that time is coming closer for them,” said Holmes. The reason for this, he added, is that there is an “increasing realisation of the difficulties that doing nothing, but saying that at some stage undefined in the future you will do something, is causing for industry and the economy”.
A recent study by Deloitte found that just one third of UK business leaders believe their company has a high or very high level of expertise in AI, while the biggest fear is being able to comply with regulations.
Generative AI lead at Deloitte UK, Lorraine Barnes, said the challenge for organisations will be keeping pace with rapid advances.
“Many business leaders are also mindful of the risks which GenAI may present, the biggest being around regulation,” she explained. “Regulations are developing quickly and organisations expect to face a level of uncertainty over their ability to meet them.”
According to Holmes, the lack of concrete rules is breeding uncertainty and instability among businesses, who are doing the “logical” thing by aligning with the EU’s AI Act in the absence of a UK one.
However, critics of Europe’s AI legislation argue that the structure of it is overly prescriptive rules make it less able to deal with the agile nature of swiftly evolving AI technology, meaning it could make innovation more challenging – an issue that the UK is trying to avoid.
As the election clock is ticking, Holmes hopes this pressure will spur both the current government into greater focus and opposing parties into making a clear positive commitment to legislating for AI in the UK.
“Everybody has seized the reality that time is short.”