Hinton Advocates for Nurturing AI Systems
Often referred to as the “AI Godfather,” Jeffrey Hinton encourages researchers to create AI systems that possess nurturing qualities. Experts in the field believe machines should be trained to care for humanity like they would for their own children.
Speaking at the recent AI4 conference in Las Vegas, Hinton suggested that the notion of keeping AI strictly under human control is impractical. Instead of a human acting as the sophisticated leader of AI, he imagines a future where humans foster ultra-responsive machines, similar to how children rely on their parents.
“We need to create machines that are smarter and care for us like we’re their babies,” Hinton stated during his presentation. He emphasized that the goal of AI development should not only focus on creating increasingly intelligent systems but also on ensuring these systems are designed with genuine concern for human well-being.
Within Hinton’s framework, humanity shifts from a position of command to one of nurturing, even if AI develops capabilities that could overshadow human abilities. He compared this dynamic to effective parenting, where nurturing figures guide the growth of children who ultimately surpass them in various abilities. Hinton believes AI research should aim to cultivate similar relationships between humans and machines.
A former Google researcher, who left the company to speak more openly about AI risks, feels that his “mother’s AI” concept could bring nations together to advance safe AI development. “Every country wants AI to evolve in ways that help its citizens and not replace them,” Hinton noted. He believes that embedding nurturing instincts into AI can pave the way for more supportive technologies.
Yann Lecun, Meta’s chief AI scientist, describes Hinton’s idea as a simplified version of safety protocols he has long advocated. He refers to it as “objectively driven AI,” where systems are designed to act strictly according to predetermined goals and values.
“This means defining AI that embodies the drives and instincts seen in humans and animals,” Lecun explained in a LinkedIn post. “In addition to motivations like empathy, there are basic rules, such as ‘Don’t harm a person’ and ‘Avoid dangerous behaviors when handling sharp objects near others.'”
