Sunak to host global AI summit amid warnings of threat to human civilisation

Prime Minister Rishi Sunak meets with Demis Hassabis, CEO DeepMind, Dario Amodei, CEO Anthropic, and Sam Altman, CEO OpenAI, in 10 Downing Street. Picture by Simon Walker / No 10 Downing Street - No 10 Downing Street/Simon Walker
Prime Minister Rishi Sunak meets with Demis Hassabis, CEO DeepMind, Dario Amodei, CEO Anthropic, and Sam Altman, CEO OpenAI, in 10 Downing Street. Picture by Simon Walker / No 10 Downing Street - No 10 Downing Street/Simon Walker

Rishi Sunak is to host a global summit in London this autumn aimed at devising international rules on artificial intelligence amid warnings it could threaten human extinction, The Telegraph can reveal.

Mr Sunak is to seek Joe Biden’s buy-in for the conference during the Prime Minister’s visit to Washington DC this week.

The Technology Secretary Chloe Smith will meanwhile meet with UK allies in Paris on Tuesday to build a united front on AI regulation ahead of the West opening up talks with China on the subject.

The UK is conducting the diplomatic blitz to accelerate cooperation on AI safety amid growing fears that without regulation the technology could pose grave risks. Experts fear that if superintelligent systems are developed with capabilities far exceeding human intelligence, humanity could lose control.

Other worries include AI being used to launch novel cyber attacks, develop devastating weapons or subvert democracy by propagating mass disinformation.

Last week, the chief executives of the world’s three most advanced AI laboratories - OpenAI, Google Deepmind and Anthropic - were among the signatories to a statement which said: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”

And at a meeting in Downing Street with the three bosses last month, Mr Sunak discussed the risks of technology, “ranging from disinformation and national security, to existential threats”.

The Prime Minister has spoken about the need to put “guardrails in place” so the development of AI is “safe and secure”.

The autumn summit is likely to bring together heads of government, top officials and AI companies from across the size spectrum. Ideas on the agenda could include creating a global AI watchdog similar to the International Atomic Energy Agency, which monitors safety standards relating to nuclear energy.

Crucially, China will be invited to attend the conference.

The Government believes that because China is one of the major global players in AI, efforts to effectively regulate the technology on an international basis will fail unless there is engagement with the country.

However, there is also considerable wariness about how China and other authoritarian states could seek to use the technology.

On Tuesday, Ms Smith will chair talks with foreign counterparts at the inaugural Organisation for Economic Co-operation and Development (OECD) Global Forum on Technology in Paris - an event which the UK Government has helped fund.

The discussions are designed to make sure that the rapid development of technology such as AI is rooted in democratic values and kept safe from hostile actors. China is not attending the forum, so the event will provide an opportunity for democratic states to build a united front on the issue of AI regulation.

‘Citizens are rightly concerned’

Countries taking part in the discussions include the USA, Japan, South Korea, Israel, Australia, New Zealand, Brazil, Chile, Norway, Turkey, Ukraine and Senegal, with the EU also participating.

A Whitehall source said: “Next week, we’re bringing like-minded governments and organisations from around the world together for discussions on some of the most defining issues of our time.

“Some of these technologies are evolving on an unprecedented scale and our citizens are rightly concerned about the pace of change.

AI has the potential to radically improve lives, but the potential for misuse by hostile, autocratic regimes is an ongoing threat that we are determined to tackle.

“With these technologies being developed across borders, it is a truly global task to ensure they are built in line with our democratic values and used responsibly.”

The source added: “That’s why Britain is showing global leadership and building an international alliance rooted in these values, securing international cash to bolster the work and proving we are truly a science and technology superpower.”

The OECD forum will discuss “responsible, values-based and rights-oriented technology” with a focus on “emerging technologies”. As well as AI, it will cover immersive technologies such as the virtual reality “metaverse”, along with engineering biology and quantum technologies.

Separately, The Telegraph understands that the Government is close to appointing a chair for a key domestic AI taskforce.

In April, the Government announced £100 million in startup funding for an expert group aimed at accelerating the UK’s sovereign AI capabilities.

The Department for Science, Innovation and Technology has narrowed down to two candidates to chair the “Foundation Model Taskforce”, with the final appointment expected to be announced in the coming weeks.

The idea of London hosting an international summit on AI regulation first appeared in The Telegraph in April, when it was proposed by the two chairs of the Commons committees on Business and Trade and Science Technology, Darren Jones and Greg Clark.