Two years to save the world warns Rishi Sunak’s AI chief

AI systems will be powerful enough to ‘kill many humans’ within two years, Rishi Sunak’s adviser on artificial intelligence warns.

Experts warn AI will be capable of killing millions within two years. Picture: iStock.

AI systems will be powerful enough to “kill many humans” within two years, Rishi Sunak’s adviser on artificial intelligence has predicted.

Matt Clifford, who is helping to set up the government’s AI taskforce, said policymakers should be prepared for deadly dangers if mankind failed to control the technology.

“In the industry we talk about near-term and long-term risks, and the near-term risks are actually pretty scary,” he said on TalkTV’s First Edition. “You can use AI today to create new recipes for bioweapons or to launch large-scale cyberattacks … these are bad things.

“You can have really very dangerous threats to humans that could kill many humans – not all humans – simply from where we’d expect models to be in two years’ time.”

Last week a statement signed by 350 AI experts – including the chief executive of OpenAI, which developed ChatGPT – said the technology could eventually lead to the extinction of humanity. Clifford, who chairs the government’s Advanced Research and Invention Agency, said this “sounds like the plot of a movie” but was a real concern that meant countries must work together to regulate AI.

“I think the letter-writers were talking about … what happens once we effectively create a new species, you know, an intelligence that is greater than humans,” he said, adding that on a “bullish” timescale we could have a computer that is more clever than a human within two years. “If we … don’t know how to control it, that’s going to create a potential for all sorts of risks.”

Clifford added that those “building the most capable systems freely admit that they don’t understand exactly how they exhibit the behaviours that they do”. Asked if this was “quite terrifying”, Clifford replied: “Absolutely.” However, he also said AI could do some good. “If it goes right … you can imagine AI curing diseases, making the economy more productive, helping us get to a carbon neutral economy.”

The White House and Downing Street have convened summits with AI leaders to get to grips with technology that is developing at pace. Sunak is keen to promote the UK as a possible hub for a global regulator, modelled on the International Atomic Energy Agency. He is also keen on the development of an international research entity akin to the particle research centre CERN.

Last month he pledged £100 million in initial funding for an AI taskforce. Clifford, who is also chief executive of Entrepreneur First, a company working with technology start-ups, was brought in to advise on its creation.

European and US regulators met in Sweden this week to find common ground on a voluntary code of conduct for AI companies. Brussels is forging ahead with the AI Act, which will designate certain types of the technology as high or low risk.

A poll of 4,000 Britons has found that almost two thirds would back laws and regulations to control AI. The survey by the Ada Lovelace Institute and Alan Turing Institute found strong support for AI in areas such as cancer screening and facial recognition at border control, but concern over its use for recruitment or creating autonomous weapons.

Not all experts believe AI is the biggest threat to humanity. Arvind Narayanan, a professor of computer science at Princeton University, said: “The history of technology to date suggests that the greatest risks come not from technology itself, but from the people who control the technology, using it to accumulate power and wealth.” He wrote of AI creators’ warnings: “We should be wary of Prometheans who want to both profit from bringing the people fire and be trusted as the firefighters.”

The Times