But what are these doomsday scenarios and how are the machines supposed to wipe out humanity?
human capacity
Most of the scenarios start from the same point: one day machines will surpass human capabilities, run out of control and refuse to be turned off.
“Once we have machines that aim for self-preservation, we’re in serious trouble,” AI scholar Yoshua Bengio said during a conference this month.
But since these machines do not yet exist, imagining how they could doom humanity is often the task of philosophy and science fiction.
The Swedish philosopher Nick Bostrom has mentioned an “intelligence explosion”, which will occur when super-intelligent machines begin to design other machines in their turn.
Bostrom’s ideas have been dismissed by many as science fiction, not least because he once argued that humanity is a computer simulation and supported theories close to eugenics.
He also recently apologized after a racist message he sent in the 1990s was exposed.
However, his thoughts on AI have been highly influential, inspiring both Elon Musk and Professor Stephen Hawking.
super intelligent machines
If super-intelligent machines are going to destroy humanity, they will surely need a physical form.
The red-eyed cyborg played by Arnold Schwarzenegger in “Terminator” has proven to be a striking image.
But experts have dismissed the idea. “This sci-fi concept is unlikely to become a reality in the next few decades, if ever,” wrote an anti-war robot activist group, Stop Killer Robots, in a 2021 report.
However, the group has warned that giving machines the power to make decisions about life and death is an existential risk.
Robotics expert Kerstin Dautenhahn, of the University of Waterloo in Canada, played down those fears.
It is unlikely, he told AFP, that AI would give machines greater reasoning abilities or instill in them the desire to kill all humans.
“Robots are not bad,” he said, though he admitted that programmers can make them do bad things.
Deadly chemicals and viruses
A less fanciful scenario is that of “villains” using AI to create new viruses and spread them.
Powerful language models like GPT-3, used to create ChatGPT, are extremely good at inventing horrible new chemical agents.
A group of scientists using AI to help discover new drugs conducted an experiment in which they tuned their AI to invent harmful molecules.
They managed to generate 40,000 potentially poisonous agents in less than six hours, according to the journal Nature Machine Intelligence.
Artificial intelligence expert Joanna Bryson of the Hertie School in Berlin explained that it is perfectly possible that someone will figure out a way to spread a poison like anthrax more quickly.
“But it’s not an existential threat. It’s just a horrible, horrible weapon.”
advanced species
Hollywood rules dictate that period disasters must be sudden, huge and dramatic, but what if the end of humanity was slow, silent and not final?
“Our species could come to an end without having a successor,” says philosopher Huw Price in a promotional video for Cambridge University’s Center for the Study of Existential Risk.
But in his opinion there are “less bleak possibilities” in which humans enhanced with advanced technology could survive.
A picture of the apocalypse is often framed in evolutionary terms.
Well-known theoretical physicist Stephen Hawking argued in 2014 that ultimately our species will no longer be able to compete with artificial intelligence machines, telling the BBC it could “signify the end of the human race”.
Geoffrey Hinton, who has spent his career building machines that resemble the human brain, speaks in similar terms of “superintelligences” that will outright outperform humans.
He recently told US broadcaster PBS that it was possible that “humanity is just a passing phase in the evolution of intelligence.”
FUENTE: With information from AFP
