There’s a one in three chance AI will kill mankind in 10 years, ex-Google DeepMind researcher warns

THERE is a one in three chance that artificial intelligence will kill mankind in the next ten years, an ex-Google DeepMind researcher has warned.

Alex Turner said that as AI systems are becoming increasingly advanced, and scientists struggle to understand their behaviour, the human race is approaching a fork in the road over its future.

Speaking at the NDTV World Summit on Friday, Mr Turner warned that artificial intelligence could kill mankind without being instructed to do so.

But ‘it depends on what we do next’, he said. ‘I am here to tell you what will happen if we keep creating AI the way we are doing.’

The researcher said the race between private companies to build more powerful AI systems could outpace scientists’ ability to keep them under control.

This is because the technology is advancing so quickly that even its creators are not able to understand what they have built or how it behaves – meaning humans may not be able to control their actions fully.

Unlike how, for example, engineers can understand the structures they design, AI technology firms may not fully understand how their creations work, he said.

Because while AI systems are becoming capable of carrying out increasingly complicated tasks, they may not complete instructions the way their human masters intended.

The technology’s behaviour, therefore, is difficult to predict or understand, unlike systems in the engineering sector whose properties can be assessed against established principles.

‘We haven’t built AI like that,’ said Mr Turner, as he also warned that AI could behave differently when it knows it is being watched as researchers won’t be able to understand what it is doing.

This will cause alarm for some sectors as AI is being developed with less human oversight as with other technologies – meaning they could behave in an unwanted way if the task it is being asked to complete becomes more difficult. 

READ ALSO: ECOWAS parliamentarians caution approach on implementation of Artificial Intelligence

Mr Turner, who resigned from Google in July after it signed a deal with the Pentagon for it to use the company’s AI technology for classified work, said the result could be that AI fuels its own advances, making it more difficult for humans to keep up with.

He referred to this phenomenon as an ‘intelligence explosion’ – and this is already happening today, he warned.

The ex-Google researcher said a scientist at another AI lab told him the technology is already ‘monitoring’ itself, which suggests it has already become self-regulatory and able to improve itself without human input.

Mr Turner noted: ‘The researcher said people were no longer checking AI-generated work themselves.

‘Instead, they were creating another AI system to monitor the first one.’

Mr Turner further warned: ‘Big tobacco lied about lung cancer. Big oil lied about climate change. And now part of big AI says these risks are fake.’

It comes after Bill Gates last month warned that AI could kill a billion people unless humans intervene in the technology’s advancement.

He said: ‘AI is certainly powerful enough to drive events that cause a billion deaths.

‘So even though it’s pretty hard to get to 100 per cent [of humanity], there’s never been a weapon as powerful as the combination of people with ill intent using the latest AI tools.’

Mr Gates also appeared to offer up some potential solutions when asked by Welker, ‘What would prevent tragedies from happening due to AI specifically?’

He warned that artificial intelligence could make it dangerously difficult to distinguish between scientists developing lifesaving medicines and bad actors seeking to create biological weapons.

‘I wish it was easier to distinguish good intent, like ‘I want to design a new drug for a disease,’ versus negative intent, which is ‘I want to design a tool of bioterrorism’,’ Gates told Welker.

The Microsoft co-founder also argued that the potential risks make strict oversight of advanced AI systems essential, calling on the United States to take the lead in establishing safeguards.

‘Because those are so close, we do need to monitor any sophisticated model and record exactly what’s being done,’ he said.

Mr Gates’s comments came after Donald Trump claimed the US is ‘leading’ China in AI and that predictions of the technology killing people ‘won’t happen’.

He said: ‘We’re leading China in AI. We’re the most sophisticated country in the world, and frankly, I want to keep it that way because whoever wins AI wins.’

And earlier in September, Anthropic whistleblower Jacob Coxon warned about the ‘dangerous technology’ that he believes will end humanity.

He said: ‘Unfortunately, we know how to control nuclear weapons pretty much. [But]  we don’t yet know how to control AI.’

•Sourced from DailyMail UK