By Jess Jones
Prime Minister Rishi Sunak has announced a “landmark agreement” which will see the UK government test the safety of new artificial intelligence (AI) models before they are released.
“Until now the only people testing the safety of new AI models have been the very companies developing it. That must change.
“So building on the G7 Hiroshima process and the global partnership on AI, like-minded governments and AI companies have today reached a landmark agreement,” Sunak said in a closing speech at the AI safety summit on Thursday.
“We will work together on testing the safety of new AI models before they are released,” he told the audience of technology leaders, politicians and companies from over the globe.
Sunak said there will be external validation of the companies, testing for national security risks, monitoring of their models during training, and before and after deployment.
“What we can’t do is expect companies to mark their own homework,” Sunak added later on.
“Ultimately that is the responsibility and knowledge of a sovereign government…no one else can do that.”
Earlier this week US president Joe Biden said organisations whose AI models could threaten national security must report on how they are ensuring their AI tools are safe.
Yesterday at Bletchley Park, where the government’s summit is being held, twenty-eight governments including the UK, US, EU and China agreed to the Bletchley declaration, named after the site of the AI safety summit – although not everybody is happy with this.
