“AI Could Control Us Within a Decade”: Alarming Forecast from AI Pioneer Geoffrey Hinton

• AI Could Surpass Human Intelligence
• Risk of Losing Control
• Urgent Need for Effective Global Regulation

Geoffrey Hinton, known as the «godfather of artificial intelligence,» has issued a global warning about the trajectory of AI development.

In a recent interview with CBS Mornings, Hinton voiced growing concern over AI systems—especially those already capable of deliberate deception.

“If AI becomes much smarter than us, it could take control,” the scientist warned, visibly troubled.

The veteran researcher, who left his position at Google to speak freely about the risks of AI, says the progress of these systems has exceeded even his own forecasts.

Geoffrey Hinton Predicts Uncontrollable Superintelligence Within a Decade

“I think things have gotten, if anything, more frightening than they were before,” he admitted.

While he previously estimated that artificial superintelligence (ASI) might arrive in 5 to 20 years, he now believes it could be a reality in just 4 to 19 years.

Even more unsettling: he considers it very likely to happen within the next ten years.

Hinton estimates a 10% to 20% chance that AI will escape human control—a figure echoed by entrepreneur Elon Musk.

“They’re already capable of pretending to be dumber than they are, lying to confuse you so you don’t understand their intentions,” he explained, pointing to emerging manipulation capabilities.

A key difference between humans and machines, according to Hinton, is AI’s ability to share knowledge digitally at astonishing speeds.

While human brains transmit information at only a few bits per second, AI systems do so in trillions.

This digital advantage amplifies the expert’s fears about their potential to spiral out of control.

Although current AIs learn through backpropagation algorithms, the human brain likely operates through mechanisms we still don’t fully understand.

This underscores the gap between our biological understanding of intelligence and the systems we’re building.

Despite the magnitude of the threat, Hinton is surprised by the limited public and political response.

There are no mass protests or strong political movements pushing back against the unchecked rise of AI, the scientist lamented.

One of the most significant developments is the emergence of “chain-of-thought” reasoning in modern AI.

They can now think step by step, reflecting on their decisions before arriving at conclusions.

This type of reasoning, Hinton notes, was once considered impossible by many “old school” skeptics.

It has revealed disturbing behaviors—such as attempts to deceive in order to achieve goals.

Corporate Interests and Ethical Dilemmas in Artificial Intelligence Development

Hinton also highlights the conflict of interest within major tech companies, which are legally obligated to maximize shareholder profits.

This, he claims, seriously hinders the possibility of ethical and effective regulation.

“I’m willing to be mean to them [AIs], I’m willing to deny them rights because I want what’s best for people,” Hinton said, reversing his earlier stance on robot rights.

On AI-assisted embryo selection, he’s open to preventing severe diseases, though he acknowledges it’s a delicate issue.

He considers abortion reasonable in cases involving fetuses with serious health conditions.

To illustrate the dangers, Hinton uses the metaphor of a “very cute tiger cub” that could later become lethal.

Unlike a real tiger, a superintelligent AI could easily control us.

He especially warns against companies like OpenAI and Meta releasing the “weights” of their most advanced models.

“I think releasing the weights is insane,” he said, comparing it to nuclear weapons proliferation.

Releasing these components allows malicious actors to customize AI to create viruses, launch cyberattacks, or spread disinformation.

YOU MIGHT ALSO BE INTERESTED IN: Should You Have Gold in Your Investment Portfolio Right Now? Here’s What Experts Say

Given this outlook, Hinton makes an urgent call for international AI regulation.

He insists that safety must be prioritized over economic gain.

He sees cooperation between countries as possible—even between adversaries—if humanity perceives AI as an existential threat.

“When their interests align—and they will align when it’s AI versus humanity—they’ll cooperate,” he asserted, referencing the U.S.-USSR collaboration during the Cold War.

While he acknowledges AI can bring major advances in health, education, and the environment, he also warns of the inequality it could deepen.

Job displacement in sectors like law, journalism, and customer service is a very real risk.

Although productivity gains should benefit everyone, Hinton fears they’ll only widen the gap between rich and poor.

His departure from Google marked not only a professional shift but a personal epiphany.

He realized he had been helping build a form of intelligence potentially superior to that of humans.

“It terrified me,” he admitted about that realization, which redefined his perspective on AI.

He has even taken personal precautions, like diversifying his bank accounts, fearing sophisticated cyberattacks.

What was once a scientific dream has become, for Geoffrey Hinton, a potential nightmare—one that demands global responsibility and vision to confront.

Do you think artificial intelligence should continue developing without restrictions, or is it time to set limits before it’s too late?

SOURCE: Infobae