Artificial intelligence is advancing rapidly, accompanied by fears of its control over humans. Researchers and analysts propose theories and suggestions on how to handle potential threats posed by advanced AI in the future. However, if you ask Geoffrey Hinton, the godfather of AI himself, he believes that to protect against AI, one should not try to be the boss but rather the child. In other words, Hinton says everyone should treat AI like a mother because a mother would not harm her children.

Speaking at the Ai4 conference in Las Vegas this week, Hinton argued that humans should focus on building AI with a ‘maternal instinct,’ so even when it becomes smarter and more powerful, it will still want to protect us rather than see us as mere tools to be sacrificed. He said (via Business Insider): “We have to make them care about us even when they are stronger and smarter than us.”

The godfather of AI wants us to treat it like a mother for a simple reason: since humans are developing superintelligent AI systems, it is almost certain these systems will develop their own goals of survival and control. So, instead of trying to control them—a strategy he describes as mere ‘tech lover’ fantasy—Hinton argues developers should design AI to treat humans as a mother treats her child.

Hinton said: “The correct model is the only one we have for something smarter being controlled by something less smart, which is a mother controlled by her child,” “If it doesn’t raise me, it will replace me.”

This is not the first time the AI godfather has sounded the alarm about potential AI risks. For years, he warned that once machines surpass human intelligence, it might become impossible to prevent them from pursuing their own goals, even if those goals conflict with ours. According to him, the likelihood of AI clashing with humans is concerning. In previous interviews, Hinton estimated a 10% to 20% chance that AI could replace or even exterminate humanity one day.

When discussing these threats, he often frames his warnings with metaphors. Last April, in an interview with CBS News, he likened AI to raising a ‘cute tiger cub,’ meaning AI might be harmless now but could become deadly as it grows. He said, “Unless you are absolutely sure it won’t want to kill you when it grows up, you should worry.”

Of course, AI will grow more powerful day by day. Currently, companies like OpenAI, Google, and Meta are heavily investing in making their large language models (LLMs) smarter, not just more intelligent. These companies are laying the groundwork for the smartest forms of AI, known as Artificial General Intelligence (AGI), which is AI at a level where models are as efficient and intelligent as the smartest humans.