BLETCHLEY PARK, England (Reuters) – British Prime Minister Rishi Sunak on Thursday said that like-minded governments and artificial intelligence companies had reached agreement on testing new AI models before they were released.
Sunak is hosting a two-day AI Safety Summit where on Wednesday he secured China’s backing for an international effort to manage the risks of the technology.
On Thursday, Sunak announced a further “landmark agreement” with “like-minded governments and AI companies.”
“We will work together on testing the safety of new AI models before they are released,” Sunak said at a press conference at the summit.
A list of the countries that had signed up to the safety testing collaboration did not include China, whose representatives were not included in the second day of talks.
Representatives from the United States, European Union, France, Germany, Italy, Japan, South Korea, Singapore, Canada and Australia were involved in the talks.
“It is critical that frontier AI is developed safely and that the potential risks of new models are rigorously assessed before and after they are deployed, including by evaluating for potentially harmful capabilities,” Britain said in a chair’s statement summary of the participants’ discussions.
Asked why China had not been invited to discuss AI model testing, Sunak said it had been right to invite China to some parts of the summit, in line with Britain’s policy to engage with China where possible but also protect against risks.
“We achieved the outcome that we wanted, which was for them to be here, to be engaged … and for them to have signed up to the Bletchley Park communique,” he said.
(Reporting by Paul Sandle, Martin Coulter and William James, writing by Alistair Smout)
