In dangerous situations, we should be able to turn off artificial intelligence

Eric Schmidt, the former CEO of Google, warns that when a computer system, such as artificial intelligence, reaches a dangerous point where it can improve itself, we should seriously consider shutting it down.

In an interview with ABC News about the development of computer systems and artificial intelligence, Eric Schmidt said: “Ultimately when you tell a computer: learn anything and do anything. “We are reaching a dangerous point: when the system can heal itself, we should seriously consider shutting it down.”

If artificial intelligence systems become too powerful, we may need a kill switch, warns former Google CEO. According to Schmidt, as systems become increasingly autonomous, they could pose new and more serious threats to humanity.

The comments of the former CEO of Google about the dangers of artificial intelligence

Schmidt predicts that artificial intelligence will eventually evolve from a single agent like Microsoft’s Copilot to a more complex system that can make decisions on its own. According to him, when artificial intelligence reaches that stage, it will be time for humans to step in and shut down the system. Also, humans must make sure that the AI ​​itself cannot deal with its shutdown.

“Metaphorically speaking, it’s better to have someone with their hand on the (AI) fork all the time,” Schmidt said.

Schmidt also looks at this advanced technology from another aspect; He told the ABC presenter that artificial intelligence will make every human have a few smart people in their pocket; Anyone on this planet could easily have an Albert Einstein or a Leonardo da Vinci for themselves to use in various tasks.

Schmidt believes that governments should begin to regulate AI to ensure that humans enjoy these benefits without suffering harm. Of course, he is not the only one who has such an opinion. Earlier, Sam Altman, CEO of OpenAI, also called on governments to form an “international agency” similar to the Atomic Energy Agency to ensure the safety of AI models.