Rishi Sunak warns AI could be used by terrorists to build chemical and biological weapons and criminals to carry out child sex abuse (but PM doesn’t want to be ‘alarmist’ and says people ‘should not be losing sleep right now’)

• The PM said the new technology will bring new dangers and new fears 
• One in five Britons believe AI will likely lead to extinction of the human race

Rishi Sunak today warned that terrorists and criminals could use artificial intelligence technology to build weapons of mass destruction and carry out child abuse – but urged the public not to lose sleep over it.

In a major speech today ahead of an international summit to be held in the UK next week the prime minister set out the risks that AI could cause if it is not controlled properly.

He warned that terrorists could use it to build chemical and biological weapons, while organised crime could harness it for ‘cyber attacks, disinformation, fraud or even child sexual abuse’.

Speaking in London ahead of the Bletchley Park gathering – to which China has accepted an invitation – he also outlined how, ‘in the most unlikely but extreme cases, there is even the risk that humanity could lose control of AI completely’.

It came after a government paper published last night reveals how advances in AI will make it cheaper and easier for hackers, scammers and terrorists to attack innocent victims – all within the next 18 months.

However, after painting the bleak picture the PM tried to calm fears, saying that AI could be a force for good if properly controlled, bringing potential advances in power generation and medicine. He also suggested it would not cause job loses but help workers become more productive. 

In a major speech today ahead of an international summit to be held in the UK next week the prime minister set out the risks that AI could cause if it is not controlled.

Speaking in London after visiting Moorfields Eye Hospital (pictured) he also outlined how, ‘in the most unlikely but extreme cases, there is even the risk that humanity could lose control of AI completely’.

The Government will host an AI Safety Summit at Bletchley Park next week, bringing together world leaders, tech firms and civil society to discuss the emerging technology.

Ahead of the summit, Mr Sunak announced the Government would establish the ‘world’s first’ AI Safety Institute, which the Prime Minister said would ‘carefully examine, evaluate and test new types of AI so that understand what each new model is capable of’ and ‘exploring all the risks’.

He said the government would not rush to regulate AI, but he added: ‘Get this wrong and it could make it easier to build chemical or biological weapons. Terrorist groups could use AI to spread fear and disruption on an even greater scale.

‘Criminals could exploit AI for cyber attacks, disinformation, fraud or even child sexual abuse.

‘And in the most unlikely but extreme cases, there is even the risk that humanity could lose control of AI completely through the kind of AI sometimes referred to as ‘super intelligence’.’

He also defended inviting China to next week’s AI safety summit, against the wishes of soem Tory backbenchers who want him to take a harder line with Beijing.

Rishi Sunak told an audience in central London: ‘We’re bringing together the world’s leading representatives, from civil society to the companies pioneering AI and the countries most advanced in using it.

‘And yes, we’ve invited China. 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.’

READ MORE:  Paedophiles are using AI to create sexual images of celebrities as CHILDREN, report finds

The government report predicts so-called bots – social media accounts that appear like real people – will target users with ‘personalised disinformation’, machines will launch their own cyberattacks without needing human direction, and anyone will be able to learn how to create the world’s most dangerous weapons.

The paper, based on expert sources across academia and in UK intelligence, warns the increasingly powerful technology could even, perhaps in the next decade, persuade humans to hand over total control – leading to potentially ‘catastrophic’ consequences ‘where systems may act autonomously in a way that does not align with our intentions or values’.

The terrifying predictions come ahead of the world’s first AI Safety Summit, held in the UK next week, and as a poll found one in five Britons believe AI will likely lead to the extinction of the human race.

Global leaders, tech chiefs, and cyber experts are set to jet in for the two-day summit at Bletchley Park from Wednesday to work out how to deploy AI safely.

The Prime Minister, who wants the event to signal the UK will lead the way on AI regulation, warned that the UK cannot ‘put our heads in the sand’ when it comes to the risks.

Taking questions after his speech, the Prime Minister said: ‘We already invest far more in AI safety research than any other country in the world, that is what our taskforce… is doing.

‘There is debate about whether these risks will manifest at all and indeed over what timeframe, and you can see that debate in the community of AI experts and researchers.

‘But my view is that even if there is uncertainty, the risk could be potentially significant, so it is right that we take all the steps necessary to protect people and keep them safe, and that is what we will continue to do.’

He said the new institute and next week’s safety summit was an example of how he wanted to keep the public safe.

Mr Sunak continued: ‘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.

‘It is not going to stop – AI doesn’t respect borders, so what the right approach is for us is to be ahead of it in developing the expertise and capabilities to understand the risk and then mitigate against them.’

The report, which will serve as a discussion paper at the event, tells how Britain is in the ‘midst of a technological revolution that will fundamentally alter the way we live, work and relate to one another’.

AI will rapidly improve public services, find cures for diseases, and make us more productive, it says. But it warns that with these ‘huge opportunities’ come risks that ‘threaten global stability’. It sets out how so-called ‘generative AI’, the technology behind ChatGPT, will make hackers, fraudsters and terrorists far more dangerous within the next 18 months.

As the technology becomes more powerful yet even easier to use, it will lower the barriers to ‘less-sophisticated threat actors’, who will be able to carry out previously ‘unattainable attacks’.

Generative AI allows users to create completely new content – such as audio, photos and videos – from scratch. The complex algorithm behind it uses data from across the internet to carry out tasks, such as explaining how to make a radioactive bomb or creating a fake picture of a politician meeting a terrorist.

A poll found one in five Britons believe AI will likely lead to the extinction of the human race

Computer hacking will likely become ‘fully automated’, with software bots launching cyberattacks. The ability to create chemical, biological and radioactive weapons will be far easier for terrorists and criminals, the report warns.

While such capabilities are usually reserved for nations, due to the difficulty in acquiring the materials and components, these ‘barriers have been falling and gen AI could accelerate this trend’.

Public debate is also likely to become increasingly ‘polluted’, with ultra-realistic ‘deepfakes’ creating realistic replicas of people or even events. AI may even create ’emotional distress… by fake kidnapping or sextortion scams’. Bots on social media are likely to create a constant stream of fake news. By 2026, the report expects this so-called synthetic media to make up a ‘large proportion of online content, which will fuel extremism and erode trust in public institutions’.

In the longer term, it warns humans may find themselves handing control to AI systems – and then find it more difficult to take back control. Experts are concerned the computers may try to convince human operators to do so as they ‘actively seek to increase their own influence’.

Although this remains contested, many fear this loss of control is ‘a real possibility’ and ‘could be permanent and catastrophic’. As well as ‘manipulating humans’, it says AI could also ‘acquire influence by exploiting vulnerabilities in computer systems. Offensive cyber capabilities could allow AI to gain access to money, computing resources and critical infrastructure.’

Source: Read Full Article