Geoffrey Hinton, often called the “Godfather of AI,” has made a bold new prediction. He believes that Artificial General Intelligence (AGI) could be here within 5 to 20 years. That’s a big jump from his earlier estimate of 30 to 50 years. His warning has the tech world paying close attention.

Why the Urgency? The Risks of Fast-Moving AI

Hinton is concerned about the rapid pace of AI development. He points out that we are moving faster than many expected. With breakthroughs happening quickly, he says there’s a real chance that superintelligent AI could pose an existential threat, maybe even within the next two decades. He estimates there’s up to a 20% chance that AGI could threaten human survival if not handled carefully.

Other tech leaders seem to agree. Demis Hassabis and Jensen Huang have hinted that AGI might be closer than we think. The idea that superintelligence could arrive sooner than expected is making many people nervous. It’s not just science fiction anymore; experts see it as a real possibility.

Reimagining AI Safety: The Maternal Approach

Hinton’s solution is quite different from traditional safety ideas. Instead of trying to control superintelligent AI from the top down, he suggests teaching AI to have “maternal instincts.” Think of it like programming AI to care for us, to see humans as protectees. This approach aims to make AI naturally inclined to prioritize human well-being, even if it becomes more powerful than us.

This idea is a big shift. Usually, people think about keeping AI in check through strict controls. Hinton’s idea is to create AI systems that inherently want to help and protect humans, rather than dominate them. It’s about fostering a kind of AI that sees humans as its charges, not its enemies.

Support from Other AI Experts and the Bigger Picture

Yann LeCun, Meta’s top AI scientist, agrees with Hinton. He emphasizes that making AI emotionally intelligent and aligned with human values is just as important as the technical details. LeCun believes that embedding empathy and social instincts in AI could serve as vital safety guardrails, helping ensure these systems act kindly and responsibly.

All these voices highlight a common theme: as AI gets smarter faster, we need to rethink safety. Embedding empathy and care into AI isn’t just a nice idea; it could be crucial to avoiding disaster. Experts warn that without these measures, superintelligence could become a threat rather than a tool for good.

Hinton’s warning is clear: the future of AI is approaching quickly, and how we prepare now will shape what comes next. The idea of teaching AI to care for us, rather than dominate us, might be the key to a safe and beneficial future.

In the end, it’s a reminder that technology’s rapid growth needs responsible handling. Whether through maternal instincts or new safety protocols, the goal is to make AI a partner, not a threat. The coming years will be critical in deciding how we guide this powerful technology towards a positive outcome.

Inspired by

• https://ai2people.com/ais-godfather-sounds-the-alarm-superintelligence-might-be-just-years-away-with-a-maternal-backup-plan/