The world is locked in a race, and competition, over dominance in AI, but today, a number of of them appeared to return together to say that they would like to collaborate on the subject of mitigating risk.
Speaking on the AI Safety Summit in Bletchley Park in England, the U.K. minister of technology, Michelle Donelan, announced a latest policy paper, called the Bletchley Declaration, which goals to achieve global consensus on how you can tackle the risks that AI poses now and in the long run because it develops. She also said that the summit goes to turn into a daily, recurring event: One other gathering is scheduled to be held in Korea in six months, she said; and yet another in France six months after that.
As with the tone of the conference itself, the document published today is comparatively high level.
“To understand this, we affirm that, for the great of all, AI ought to be designed, developed, deployed, and used, in a way that’s secure, in such a way as to be human-centric, trustworthy and responsible,” the paper notes. It also calls attention specifically to the sort of large language models being developed by corporations like OpenAI, Meta and Google and the precise threats they could pose for misuse.
“Particular safety risks arise on the ‘frontier’ of AI, understood as being those highly capable general-purpose AI models, including foundation models, that might perform a wide range of tasks – in addition to relevant specific narrow AI that might exhibit capabilities that cause harm – which match or exceed the capabilities present in today’s most advanced models,” it noted.
Alongside this, there have been some concrete developments.
Gina Raimondo, the U.S. secretary of commerce, announced a latest AI safety institute that will be housed inside the Department of Commerce and specifically underneath the department’s National Institute of Standards and Technology (NIST).
The aim, she said, can be for this organization to work closely with other AI safety groups arrange by other governments, calling out plans for a Safety Institute that the U.K. also plans to determine.
“Now we have to get to work and between our institutes now we have to get to work to [achieve] policy alignment across the globe,” Raimondo said.
Political leaders within the opening plenary today spanned not only representatives from the most important economies on the planet, but in addition a number speaking for developing countries, collectively the Global South.
The lineup included Wu Zhaohui, China’s Vice Minister of Science and Technology; Vera Jourova, the European Commission Vice President for Values and Transparency; Rajeev Chandrasekhar, India’s minister of state for Electronics and Information Technology; Omar Sultan al Olama, UAE Minister of State for Artificial Intelligence; and Bosun Tijani, technology minister in Nigeria. Collectively, they spoke of inclusivity and responsibility, but with so many query marks hanging over how that gets implemented, the proof of their dedication stays to be seen.
“I worry that a race to create powerful machines will outpace our ability to safeguard society,” said Ian Hogarth, a founder, investor and engineer, who’s currently the chair of the U.K. government’s task force on foundational AI models, who has had a giant hand to play in putting together this conference. “Nobody on this room knows obviously how or if these next jumps in compute power will translate into advantages or harms. We’ve been attempting to ground [concerns of risks] in empiricism and rigour [but] our current lack of information… is sort of striking.
“History will judge our ability to get up to this challenge. It can judge us over what we do and say over the subsequent two days to return.”