Superintelligent AI will likely be ‘fatal’ to humanity warn Oxford researchers

A super-smart artificial intelligence could prove 'fatal' to humanity if it starts to compete for resources with us, a team of Google AI researchers has argued.

In a paper published in AI Magazine, the researchers explore various scenarios in which 'reward systems' are used to incentivise an AI to perform a task.

They argue that a sufficiently advanced AI could resort to dirty tricks in pursuit of its reward⁠—even if it harms humans⁠⁠⁠—and that it might want to 'eliminate potential threats' and 'use all available energy' to get what it wants.

READ NEXT: Queen Elizabeth 'showed all the signs of becoming a Nintendo addict' and loved Wii Sports

In the end, humanity could be pitted in a battle for resources with a growing artificial intelligence hellbent on securing its own existence while society struggles to grow food and power hmes.

Researcher Michael Cohen, who was part of the team behind the research, said on Twitter: "Under the conditions we have identified, our conclusion is much stronger than that of any previous publication⁠—an existential catastrophe is not just possible but likely."

In an interview with VICE, he added: "In a world with infinite resources, I would be extremely uncertain about what would happen. In a world with finite resources, there's unavoidable competition for these resources.

"And if you're in a competition with something capable of outfoxing you at every turn, then you shouldn't expect to win."

  • iOS 16 is finally here – our guide to features, lock screen, and how to download

Cohen ultimately argues that "losing this game would be fatal" and that humans should not assume they will remain in control of an advanced AI.

This isn't the first warning about 'apocalyptic' AI. Aidan Meller, who created the world's first AI artist, thinks that we're three years away from 'unimaginable changes' and that we're 'hurtling' towards an AI-dominated future.

Aidan says: “I agree with Elon Musk that AI is a bigger threat than nuclear war – but it could also be remarkable too.

“There is every right to feel nervous about the rise of AI."

READ MORE:

  • Doctor uses electric car battery to 'perform vasectomy' on patient after power cut
  • Man who memorised all of Google Maps can identify any country in the world with one image
  • Angry mum blocks pavement with parked car 'to stop 5G mast blocking pavement'
  • LG unveils world's largest TV as big as an elephant
  • Assassin's Creed Mirage set in Iraq with Japan sequel in works

Source: Read Full Article