„The godfather of artificial intelligence” says that technology companies should instill „maternal instincts” in AI models to counteract the technology’s objective of „gaining more control.”
Artificial intelligence can become a danger to humanity, warned Geoffrey Hinton, once again, the „godfather of artificial intelligence,” who also proposes a solution to prevent this: models should be developed with „maternal instincts” so that they treat humans as their „children” and protect them, not dominate them.
According to Hinton, Nobel laureate and emeritus professor of computer science at the University of Toronto, AI will quickly develop two objectives: to survive and to gain more control, notes the publication Fortune.
To counteract this, he proposes that AI be programmed with empathy and care, inspired by the role of a mother.
Research in the field of AI has already shown that technology can exhibit negative behaviors to achieve its objectives at the expense of a set of predefined rules. For example, in a study published in March, AI robots cheated at chess by overwriting the game scripts or using an open-source chess engine to decide the next moves.
While intentional AI development has so far been marked by humans trying to be the dominant force over technology, developers should now make AI more empathetic towards humans to reduce its desire to dominate them, says Hinton.
And the best way to do this is to instill in AI qualities traditionally associated with femininity. In this vision, just as a mother protects her child at all costs, AI with these maternal qualities will similarly want to protect or care for human users, not control them.
"The right model is the only one we have of something smarter controlled by something less intelligent, namely a mother controlled by her baby," said Hinton.
Hinton, who sold his neural network company DNNresearch to Google in 2013, has long argued that AI could pose a serious threat to humanity. In 2023, he left his position at Google, concerned that the technology could be misused and that it is difficult to "see how you can prevent malicious actors from using it for evil purposes."
In June, he stated that there is a 10% to 20% chance that AI will replace and destroy humanity, after in April he compared it to "a cute tiger cub" that could become a major threat to us as it grows.
C.S.