Artificial intelligence (AI) could exacerbate security risks, including chemical and biological weapon attacks, a government report has warned.
The paper, published by the Government Office for Science, draws on sources, including intelligence assessments.
It confirmed that AI misuse could lead to loss of control over the technology, which could be “permanent and catastrophic.”
Handing over control of important decisions to AI systems can result in increased consumption of extremist content and dangerous medical misdiagnoses, the report said.
“As a result, AI systems may increasingly steer society in a direction that is at odds with its long-term interests,” it added.
In his speech, the prime minister acknowledged the risks posed by AI “superintelligence” but said it was an unlikely and an extreme example of human loss of control over AI.
Mr. Sunak said he didn’t want to be alarmist and that it wasn’t a “risk that people need to be losing sleep over right now.”
“What we can’t do is just put our heads in the sand and think this can be either be ignored or will stop happening,” Mr. Sunak added.
The discussion paper is part of the government’s plan to seize “the enormous opportunities” AI brings and to harness its benefits, while mitigating risks to public safety.
The summit will focus on certain types of AI systems based on the risks they may pose. This includes frontier AI applications, which underpin the widely known ChatGPT, and narrow AI systems, which refer to systems specified to handle a singular or limited task.
At the summit, the prime minister wants to establish a global expert panel, nominated by the countries and organisations attending. The panel is meant to publish a “state of AI science” report.
He added that Britain is “well-placed” to create the world’s first AI safety institute after AI companies “trusted the UK with privileged access to their models.”
Attendance and Ambition
Among those invited to the summit is China, whose attendance may cause some controversy, amid tense relations between London and Beijing.“I know there are some who will say they should have been excluded, but there can be no serious strategy for AI without at least trying to engage all of the world’s leading AI powers. That might not have been the easy thing to do, but it was the right thing to do,” the PM said during his speech.
Deputy Prime Minister Oliver Dowden said the ministers will wait and see whether the Chinese delegation actually arrives for the summit next week.
Among other concerns of AI development is the fear that it will automate jobs, pushing out people of their jobs.
Addressing this during his speech, the prime minister said that while all technologies change the labour market, a better way to think about AI is as a “co-pilot.”
He said that the best way to labour market security is to offer “a world-class education system.”
There are “stark differences in governance and regulatory approaches” between the EU and U.S., which the UK would struggle to reconcile, said Yasmin Afina, a research fellow at Chatham House’s Digital Society Initiative.
She added that none of the world’s most pioneering AI firms was based in the UK and the Britain should instead focus on promoting responsible behaviour in the research and deployment of AI.
