AI experts and politicians from around the world have gathered in the UK for a summit on the risks of AI, which opened with representatives from the US, China, the EU, and the UK promising to collaborate, in support of a "Bletchley Declaration" on AI safety.
The meeting at the historic Bletchley Park code-breaking facility includes politicians from many nations and will be addressed by scientists and technologists, before closing tomorrow with a live-streamed discussion between UK Prime Minister Rishi Sunak and Elon Musk at the site, where Alan Turing led a gathering of UK scientists to break German codes during World War II.
Technology Secretary Michelle Donelan opened with a call for collaboration amongst nations to mitigate risks from AI, which "compels" us to work together to make AI safe, and secure prosperity from the technology. She promised the debate would help create "guardrails" to enable safe development of the technology,
"AI is not some natural phenomenon happening to us, it is a product of human ingenuity, which we have the power to shape," said Donellan. "This is no time to bury our heads in the sand."
The US has already put a stake in the ground on AI safety, with an Executive Order on AI safety from President Biden earlier this week. In the initial panel session at Bletchley, US Secretary of State for Commerce Gina Raimondo, announced an AI safety institute to be set up at the NIST standards body in the US, promising that this would work closely with other countries including the UK.
Chinese computer scientist Wu Zhaohui told the event that leaps in technology would bring benefits, but warned that lack of transparency and other issues would pose "challenges in ethics". He said the Chinese Government wanted "open platforms" and had released principles for AI governance within the third Belt and Road Forum.
European Commission Vice-President for Values and Transparency, Věra Jourová, told the summit that the EU has already published a policy on AI in 2020, which will be updated with a new chapter on Generative AI, which could be available on 6 December.
"We need smart regulations combined with voluntary commitments," she said, calling on nations to sign up to the G7 nations' voluntary code which the EU endorsed yesterday.
The chair of the UK's Frontier AI group Ian Hogarth said he "worried" that our ability to create advanced systems would outpace our ability to manage the risks. Our current lack of understanding in the face of the challenge, he said, was "quite striking."
The summit is expected to look at risks to public safety, to the dangers of disinformation, and the potential for AI to destabilize society. Some critics have said it should look more closely at immediate risks, such as increasing discrimination and dangers to jobs.
Jourová said this was "a false debate," saying that "we need to address the risks we face today," and advising that all rules and guidance we set up now "will help us deal with more uncertain risks going forward."
The UK Government said that AI already employs over 50,000 people in the UK and contributes £3.7 billion to the country's economy each year.
A second summit is planned in South Korea in six months, followed by one in France in one year.