AI could create alien civilisation where computers rule & humans are condemned to death – tech bosses are right to worry
IMAGINE you are about to board a plane for a dream holiday when its engineers warn there is a fair chance it will crash, suggests Skype founder Jaan Tallinn.
“The sensible response is . . . don’t fly,” concludes Tallinn, who now runs The Centre for Study of Existential Risk.
The IT wizard was talking about super-fast Artificial Intelligence, which could improve life for everyone on the planet — or condemn us all to certain death.
His concern is not just for the wary passenger preparing for take-off, but for the entire human race, already airborne and cruising towards the point of no return.
Even tech giants leading the charge in multitrillion-dollar robot technology race are worried.
“I think people should be happy that we are a little bit scared of this,” said Sam Altman, whose Chat GPT is taking the world by storm.
READ MORE FROM TREVOR KAVANAGH
Eliezer Yudkowsky, of the Machine Intelligence Research Institute in Berkeley, California, thinks we should be scared witless.
“Many researchers steeped in these issues, including myself, expect that the most likely result of a superhumanly intelligent AI is that literally everyone on Earth will die.
“Not that this might be possible, but that it would be the obvious thing to happen.”
AI is with us already. It has huge potential for good, especially health care, climate change and feeding the world.
Most read in The Sun
Chatbots write poetry or music, school essays or entire books in the style of famous authors.
But there is a downside.
Facial recognition, deep fake propaganda, malicious algorithms and killer drones programmed to pick their own targets are already a sinister fact of life.
Already too late
And that’s all before AI begins to develop a mind of its own, deciding which man-made data to obey and which to dismiss — regardless of the cost in human life.
It has already proved capable, left to its own devices, of designing bio-toxins capable of wiping out populations.
Which explains the clamour from such tech geniuses as Tesla’s Elon Musk and Apple’s Steve Wozniak for an emergency halt while the world works out how to proceed safely.
Eliezer Yudkowsky thinks it’s already too late.
“Behind the outward appearance of an AI that talks to you and answers your questions are giant arrays of inscrutable numbers,” he says. “Nobody knows what sort of thinking is taking place within them.
“In our current state of ignorance, the most likely outcome is that we create an AI that does not do what we want and does not care for us or life in general.
“Such an AI will neither love nor hate us — but it will recognise that we are made of atoms that it can use for something else.
“Visualise an alien civil-isation, thinking at millions of times the speed of humans and operating in a world of creatures that are, from its perspective, very stupid and very slow.”
There is also the age-old problem of “rubbish-in, rubbish out”.
Google’s Bard recently betrayed its left-wing bias by condemning Brexit as a “bad idea", while suggesting Jeremy Corbyn had “the potential to be a great leader”.
As a dutiful investigative journalist, I interviewed Chat GPT about the hazards of AI.
It replied instantly, at fast conversation speed: “One of the most significant risks is the potential to harm human life, to malfunction or be programmed incorrectly.
“Another risk is for it to be used for malicious purposes, to perform tasks that are harmful to human life, such as developing biological weapons or conducting cyber attacks.
“The main concern is that AI systems could potentially surpass human intelligence and become superintelligent. AI systems could take over the world and potentially pose an existential threat to humanity.”
READ MORE SUN STORIES
Asked if totalitarian regimes were likely to use AI for the benefit of mankind, it replied: “It is difficult to trust totalitarian regimes such as China and Russia to develop and use AI in a responsible and beneficial way, given their track record on human rights and lack of transparency.”
Thanks to Chat GPT, you have it straight from the Trojan horse’s mouth.
A breath of fresh air
BLUNT-speaking Suella Braverman is a breath of fresh air.
Vulnerable young white girls are still being drugged, raped and trafficked by mainly British Pakistani gangs because police and Labour-controlled councils fear being branded “racist”.
Even Labour’s Lisa Nandy could not deny that, when asked by the BBC.
And what about illegal migrants who crossed The Channel “by mistake”, asked Sky’s testy Sophy Ridge.
“I don’t think paying people smugglers thousands, risking your life in a flimsy dinghy and travelling at night in freezing conditions is ‘making a ‘mistake’,” replied the Home Secretary.
More! More!