AI Minister Warns Regulation Must Move Faster Than Online Safety Bill
Rishi Sunak delivered a major speech on the UK's plans to regulate artificial intelligence ahead of the AI safety summit next week (Alamy)
The minister for AI has insisted the government must move more quickly to draw up AI safety regulation than it did to pass online safety legislation or there is a risk the technology will have “moved on”.
Viscount Camrose said the Online Safety Bill, which finally received Royal Assent this week having been initially devised when former prime minister David Cameron was still in office, cannot be the model for regulating AI.
He said governments must be “agile” in their approach due to the speed at which AI technology is moving, and “produce very, very targeted regulation” to fix problems as they arise, rather than waiting for a catch-all set of rules somewhere down the line.
Speaking to PoliticsHome podcast The Rundown, the minister in the department for science, innovation and technology, said policymakers must be “very focused on measurable, observable evidence, as opposed to oh, we are concerned that this might happen, therefore we must regulate against this”.
Ahead of the government’s landmark AI safety summit at Bletchley Park next week, when the UK is hoping to draw together the world’s biggest investors in the technology as well as world leaders to decide on how to regulate the industry, Prime Minister Rishi Sunak gave a speech setting out the long-terms risks it might pose.
He warned that AI could help make it easier for terrorists to build chemical and biological weapons, and in a worst-case scenario we could lose all control over the technology. Sunak insisted that we must not "put our heads in the sand" over these potential threats.
The Prime Minister also announced that he was setting up an AI safety institute, which Camrose agreed was needed as “the bulk of all AI safety research has been done historically, by the AI labs themselves”, and that they cannot go on “marking their homework”.
But he said the rise of AI technology was not all “doom and gloom”, pointing to the huge investment opportunities it presents, which Britain can help harness by putting itself at the forefront of regulation.
The minister cited the Online Safety Bill, which endured a torturous process from conception to the statute book as governments changed hands and different vested interests saw its focus re-shaped multiple times before it finally passed through Parliament last month.
“We cannot do that – things will have moved on a little bit in six years,” he said.
“The idea would be to, when necessary, and when driven by the evidence, produce very, very targeted regulation to fix that, and to be ready to keep doing that."
He likened some of the public anxiety over the potential damage AI could do to the fears expressed about genetically modified crops a generation ago.
“I think the debate there possibly in retrospect was wrong, in that it was one side going no, don't worry, it's going to be fine, and everybody else was saying well, we don't think it's fine, we don't trust this”, he said.
“In the end, the arguably wrong decision was made to greatly limit the use of GM foods that actually 20 years on turned out to be much safer than everybody thought.
“Let's just really have an open conversation. There are risks, but we are really taking every conceivable responsible step to deal with those risks.
“But let's be open about that, then we can all take advantage of AI.”
PoliticsHome provides the most comprehensive coverage of UK politics anywhere on the web, offering high quality original reporting and analysis: Subscribe