Rishi Sunak is worried about AI. The UK Prime Minister believes that bioterrorists will exploit the technology – which is improving rapidly – before governments can catch up.
Sunak says the next year will be crucial.
According to The Telegraph:
The Prime Minister believes there is a “small window” to address safety concerns before a powerful next generation of systems is released, Whitehall sources said.
It comes as Mr Sunak prepares to host an international AI safety summit in November, when world leaders are expected to press the top AI labs to make safety commitments as they develop new versions of their models.
Officials fear the next wave of AI could be orders of magnitude more powerful than today’s technology such as ChatGPT and that the next 12 months are crucial to getting ahead of the problem.
On Monday, the Government warned that AI could be used to create bioweapons and that humans could lose control of the software.
The Department for Science, Innovation and Technology said in a statement: “We are at a crossroads in human history and to turn the other way would be a monumental missed opportunity for mankind.”
It said the summit would focus on cases “where a bad actor is aided by new AI capabilities in biological or cyber-attacks, and loss of control risks that could emerge from advanced systems that we would seek to be aligned with our values and intentions.”
Michelle Donelan, the Science and Technology Secretary, said that she hoped AI companies would commit to a “smoke alarm” that would halt development of the models when they identify dangerous activity.
She said: “I think what we will see coming out of the summit is definitely some agreement across nations and some agreement working with companies as well.
“We need almost like a smoke alarm established so that not only are companies searching for, but have a response to the risk and we know at what level they respond and how they respond.”
However, it is understood the summit will be focused on voluntary commitments, rather than an international agreement that would require AI labs to abide by safety standards.
