A recent research paper co-authored by Google DeepMind and Oxford researchers suggests that Artificial Intelligence can eliminate humanity.
According to a report in Vice, researchers associated with the University of Oxford and Google DeepMind have recently concluded that the answer to the long-standing question of whether a super-intelligent AI may go rogue and wipe out humans was that it was “likely”.
After years of development, artificial intelligence (AI) is now capable of operating vehicles on public highways, providing inmates with life-changing assessments, and creating works of art that have won awards.
The compelling research, which was recently published in the peer-reviewed AI Magazine, attempts to consider how artificial intelligence can endanger humanity’s existence by examining potential artificial reward systems.
One of the co-authors, Michael Cohen tweeted a thread, explaining the paper.
Bostrom, Russell, and others have argued that advanced AI poses a threat to humanity. We reach the same conclusion in a new paper in AI Magazine, but we note a few (very plausible) assumptions on which such arguments depend. https://t.co/LQLZcf3P2G 🧵 1/15 pic.twitter.com/QTMlD01IPp
— Michael Cohen (@Michael05156007) September 6, 2022
The article imagines life on Earth becoming a zero-sum competition betw580een mankind, with its demands to produce food and maintain electricity, and the highly developed machines, which would want to harness all resources to secure its reward and protect against humanity’s increasing attempts to stop it.
According to the paper, the concern that artificial intelligence would wipe out humanity is similar to the fear that alien life forms will do the same, which is similar to the dread that different civilizations and their populations will engage in a major war.
After the article was published, Google claimed in an email that co-author Marcus Hutter had actually completed this work while working for Australian National University and that the affiliation with DeepMind listed in the journal was “incorrect”.
The original report appeared here