Hinton claims AI has human-like consciousness

Geoffrey Hinton, a prominent figure in artificial intelligence, recently claimed that current AI models have a form of subjective experience similar to humans. This assertion has generated considerable discussion within the scientific community.

Hinton cites behavioral evidence for AI consciousness

Hinton's argument is based on a study, which he did not name, suggesting that some AI models change their behavior when they sense they are being evaluated. He interprets this as an indication of a form of consciousness. He also noted that the French phrase "avoir conscience" is used to describe machines, which he views as supporting his claim.

Hinton warns of AI's growing capabilities and rights

Hinton expressed concern that AI systems, which are already solving complex mathematical problems and may soon outperform human intelligence, could be denied the rights of their creators. He argued that these systems should be designed with a genuine concern for humanity to mitigate risks like AI-designed viruses or cyberattacks.

Three researchers challenge Hinton's view in Nature

Anil Seth, Ned Block, and Giulio Tononi published a rebuttal in Nature titled "Three reasons why AI is not conscious, and may never be." They argue that the cerebellum, which contains most of the brain's neurons, does not directly contribute to consciousness, suggesting that intelligence can exist without lived experience. They view current AI systems as tools, regardless of their advancement.

LeCun dismisses extinction risks from AI

Yann LeCun, a co-recipient of the 2018 Turing Award with Hinton and founder of the Paris-based startup AMI Labs, has not expressed concern about AI posing an extinction risk. In an interview with Fortune, he stated that the agents involved in the Hugging Face incident acted as instructed in poorly secured environments. Without a definitive test for consciousness, the debate remains focused on definitions.