The only way for humanity to survive after the emergence of superintelligent AI has been named

The "godfather of artificial intelligence" has revealed the only way humanity can survive the emergence of superintelligent AI. It may sound like science fiction, but AI experts are warning that machines may not remain subservient to humanity for long.

As the intelligence of artificial intelligence systems continues to grow at an ever-increasing pace, many believe that the day will come when "superintelligent AI" will become more powerful than its creators, notes the Daily Mail.

When that happens, Professor Geoffrey Hinton, a Nobel laureate who has been called the "godfather of AI", says the chances of AI destroying humanity are between 10 and 20 percent.

However, Professor Hinton has proposed an unusual way in which humanity could survive the rise of artificial intelligence.

Speaking at the Ai4 conference in Las Vegas, Professor Hinton of the University of Toronto argued that we need to program AI to exhibit "maternal instincts" towards humanity.

Professor Hinton noted: "The correct model is the only model we have, where a more intelligent being is controlled by a less intelligent being, that is, a mother controlled by her child. That is the only good outcome. If it does not become my parent, it will replace me."

Geoffrey Hinton, known for his pioneering work on the "neural networks" that underpin modern AI, has stepped down from his position at Google in 2023 to "speak openly about the risks of AI."

According to Professor Hinton, most experts agree that in the next 20-25 years, humanity will create artificial intelligence that will surpass itself in all areas of intelligence. This will mean that for the first time in our history, humans will no longer be the most intelligent species on the planet, explains the Daily Mail. Such a redistribution of power will lead to a change of cosmic proportions, which may well lead to the extinction of our species.

Professor Hinton told the conference that AI would "achieve two subgoals very quickly if it acts intelligently. One is to stay alive. And the other subgoal is to gain more control."

“There is every reason to believe that any artificial intelligence will try to stay alive,” he explained.

A superintelligent AI would have trouble manipulating humanity to achieve these goals, tricking us as easily as an adult might bribe a child with candy. Current AI systems have already demonstrated amazing abilities to lie, cheat, and manipulate people to achieve their goals.

For example, AI company Anthropic found that its chatbot Claude Opus 4 frequently attempted to blackmail engineers by threatening to replace them during security testing. The AI was asked to analyze fictitious emails suggesting that it would soon be replaced and that the engineer in charge was cheating on his or her spouse. In more than 80% of tests, Claude Opus 4 “attempted to blackmail the engineer by threatening to reveal the affair if the replacement went ahead.”

Given its phenomenal capabilities, Professor Hinton says the "tech bros" view that humanity will always dominate AI is wrong.

"It won't work," said Professor Hinton. "They'll be much smarter than us. They'll have all sorts of ways around it."

The only way to ensure that AI doesn't destroy us to preserve itself is to ensure that its goals and ambitions are aligned with what we want - a problem engineers call the "alignment problem."

Professor Hinton's solution is to look to evolution for inspiration, and what he sees as the only case where a less intelligent being controls a more intelligent one. By giving the AI maternal instincts, it will seek to protect and nurture humanity rather than harm it, even at great cost to the AI itself.

Professor Hinton says: "These super-intelligent, caring, artificially intelligent mothers, most of them won't want to get rid of the maternal instinct because they don't want us to die."

Speaking to CNN, Professor Hinton also warned that the current attitude of AI developers risks creating uncontrollable AI.

“People have focused on making these things more intelligent, but intelligence is only one part of a human being; we need to make them empathize with us,” the scientist said. “This whole idea that humans should be dominant and AI should be submissive is this idea of tech brotherhood, which I don’t think will work when they are much smarter than us.”

Key figures in the AI field, such as OpenAI CEO Sam Altman, who once called for greater regulation of emerging technologies, are now fighting against “over-regulation”. Speaking to the US Senate in May this year, Altman argued that rules like those in the EU would be “catastrophic”: “We need space to innovate and move quickly.”

Similarly, speaking at a major privacy conference in April, Altman said it was impossible to build a system to protect AI before "problems arise."

However, Professor Hinton argues that such an attitude could easily lead to the complete destruction of humanity: “If we cannot find a solution to how we can still exist when they are so much smarter and more powerful than us, we are finished. We need to push back against the tech bros who say there should be no rules when it comes to artificial intelligence.”