Human extinction risk from AI on same scale as pandemics or nuclear war, Sunak warns

AI could help terrorists build bio-weapons, says PM in major speech on benefits and threats posed

Rishi Sunak has said mitigating the risk of human extinction because of AI should be a global priority alongside pandemics and nuclear war.

AI will pose major security risks to the UK within two years and has the potential to “enhance” terrorist capabilities, new analysis by the Sunak government has warned.

The technology could aid terror groups’ capabilities – including propaganda, bioweapons and attack planning – according to a report released as the PM made a major speech on the “breathtaking” pace of change.

The official report by the Government Office for Science also warned that, by 2025, AI is expected to increase the frequency and sophistication of cyberattacks, scams, fraud and other crimes.

Speaking at the Royal Society on Thursday morning, Mr Sunak said the changes brought by AI could be as “far-reaching” as the industrial revolution, or the arrival of electricity or the internet.

The PM said he agreed with experts who believe the extinction threat from AI should be treated like the threat of pandemics and nuclear war, as he called for a global expert panel to address the issue.

Mr Sunak told a press conference that AI would bring “the chance to solve problems we once thought beyond us – but it also brings new dangers and new fears”.

The PM also claimed that the UK was doing “far more than any other country to keep you safe”, as he announced that the government will establish the world’s first AI safety institute in the UK.

Quoting a statement made by hundreds of AI experts earlier this year, Mr Sunak said: “Mitigating the risk of extinction by AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”

Rishi Sunak delivers a speech on AI at the Royal Society (PA)

On the major risks ahead, the PM said: “Get this wrong and AI could make it easier to build chemical or biological weapons. Terrorist groups could use AI to spread fear and destruction on an even greater scale.”

He added: “Criminals could exploit AI for cyberattacks, fraud or even child sexual abuse … there is even the risk humanity could lose control of AI completely through the kind of AI sometimes referred to as super-intelligence.”

Mr Sunak said he did not want to be “alarmist” and highlighted the benefits AI could bring – praising its potential use in the NHS to diagnose and prevent strokes and heart attacks, as he announced an extra £100m for AI treatments for previously incurable diseases.

The PM also confirmed that he had invited China to his global AI summit at Bletchley Park next week – arguing that was “the right thing to do” because the world’s leading AI powers should discuss the threats and opportunities.

The Tory leader said it was “not an easy decision” to invite Chinese officials – and admitted that he could not say with 100 per cent certainty that they will turn up.

Deputy PM Oliver Dowden said China had accepted an invitation. He told BBC Radio 4’s Today programme: “It is the case that they have accepted but … It is the case that you wait and see who actually turns up at these events. We do expect them to come.”

The decision to invite Beijing has caused controversy in some quarters, including among Tory MPs.

Mr Sunak’s predecessor Liz Truss immediately slammed the move, saying she was “deeply disturbed” and called on the prime minister to cancel the invitation. The ex-PM hit out at the Chinese state which she said had “used and abused technology to aid its oppression of millions and attacks on freedom and democracy”.

The massive popularity of services like ChatGPT is an example of what Sunak refers to as a ‘breathtaking’ pace of change (AFP/Getty)

It comes as the White House revealed that US vice-president Kamala Harris will attend the UK summit at Bletchley Park.

Before the US vice president attends the summit, she will deliver a speech outlining her administration’s approach to AI next Wednesday, the day before the event kicks off.

Asked if the Biden administration could distract from the UK conference, Mr Sunak’s official spokesman. “It’s right that we work collaboratively with the US on this … that is entirely right.”

Meanwhile, the new UK government report warns that AI is likely to make cyberattacks, faster, more effective and larger scale. This will happen via more tailored phishing methods or by replicating malware. Another risk is what is termed the “erosion of trust in information”.

“Deepfakes”, where fake videos are created using someone’s likeness and what are termed “hyper-realistic bots”, a form of fake social media profile risk “creating fake news, personalised disinformation, manipulating financial markets and undermining the criminal justice system”, the report warns.

The Government Office for Science documents, collated using sources including UK intelligence, even says a wider threat to the future of humanity from AI cannot be ruled out.

One warns: “There is insufficient evidence to rule out that highly capable future frontier AI systems, if misaligned or inadequately controlled, could pose an existential threat.”

It also warns that AI could disrupt the labour market by displacing human workers. It suggests a so-called “robot tax” – a levy on businesses profiting from the replacement of workers by AI – might be needed to protect the economy.

Peter Kyle, Labour’s shadow science secretary, urged Mr Sunak to “back up his words” following the speech. “We are still yet to see concrete proposals on how the government is going to regulate the most powerful AI models,” he said.

Technology secretary Michelle Donelan said that by publishing the report the UK is the “first country in the world to formally summarise the risks presented by this powerful technology”. She said: “We cannot harness its benefits without also tackling the risks”.