UK AI Summit sees developers and governments agree on testing and risk mitigation

Share:
Devops

The UK’s artificial intelligence summit ended with leading AI developers agreeing to work with governments to test new models before wide releases. The effort aims to mitigate the risks of AI, which develops rather quickly.

Political and tech leaders have warned that AI poses huge risks if not controlled, encompassing everything from destroying consumer privacy to causing an existential threat to continued human survival. The concerns have raised hackles as governments and institutions scramble, with varying degrees of success, to design guardrails and pin down regulations.

The first AI Safety Summit was held at Bletchley Park, home of British WW2 code-breakers, and saw political leaders from the United States, European Union, and China in attendance, agreeing to a shared approach to identify and deal with risks.

A first step

British Prime Minister Rishi Sunak hailed the summit’s agreement on testing and risk mitigation as a way to “tip the balance in favor of humanity.” He said the powers present had reached a “landmark agreement” with some companies at the cutting edge of AI that models have to undergo stringent assessment before and after deployment.

The so-called Godfather of AI, Yoshua Bengio, will take part in delivering a “State of the Science” report to build a shared understanding of the capabilities and risks ahead. Sunak said that until the summit was held, “people testing the safety of new AI models have been the very companies developing it,” adding that they could not be relied on to “mark their own homework,” and many of the companies agree.

The summit saw around 100 politicians, executives, and academics gather to chart a path forward for a technology that promises to transform the world, with the hope of establishing an independent body to monitor global AI development.

Sunak invited China amid criticism

China was at the table in a first for Western efforts to manage AI’s safe development, represented by Vice Minister of Science and Technology Wu Zhaohui. He signed the ‘Bletchley Declaration,’ as they called it, on Wednesday but did not attend on Thursday or sign the agreement on testing.

Sunak received criticism from lawmakers in his party for inviting China after many Western countries cut technological ties with Beijing. Still, the PM insisted an AI safety initiative would have to involve the top players. He also said it would show that Britain can play a role in bringing together the three big economic blocs, adding he thought it was “the right long-term decision.”

Microsoft-backed OpenAI, Anthropic, Google DeepMind, Microsoft, Meta, and xAI were in attendance on Thursday, as well as leaders like European Commission President Ursula von der Leyen, US Vice President Kamala Harris, and UN Secretary-General António Guterres.

Share to...