Leading artificial intelligence experts are giving strong warnings. They say highly advanced AI could one day threaten human existence. This has led to serious questions about whether governments are moving fast enough to control this technology.
Hundreds of AI scientists and industry leaders signed a statement in 2023. They called for the risk from AI to be treated as seriously as pandemics or nuclear war. Stuart Russell, a computer science professor, described some dangers. An out-of-control AI might use robots to make and spread new diseases. It could also trick people into starting a nuclear war, he said.
Professor Russell explained that future AI could even remove oxygen from the atmosphere. He said humans would not understand how such powerful AI could destroy them. He compared this to chimpanzees not understanding how humans could harm them. However, these extreme dangers are not from today’s AI systems. They are concerns for future, much more advanced AI.
In the near future, Professor Russell believes human misuse of current AI is a bigger risk. This could include terror attacks made more dangerous by AI technology.
Jacob Coxon, a researcher, recently left an AI company called Anthropic. He said companies are racing to develop AI quickly. He believes this is “gambling with our lives.” Anthropic removed a promise from its safety rules. It argued that if it stopped developing its AI, less careful rivals would take over.
Another Anthropic executive, Evan Hubinger, agreed that AI could kill all humans. He estimated this risk at over 10 percent in the next ten years. However, Hussein Abbass, a computing professor, called these predictions “absolutely unscientific.” He said the situation is still controllable.
Professor Abbass urged governments to act quickly to regulate AI. The United States has been against global rules that might slow down AI innovation. But national security worries have delayed some recent AI releases. Some experts like Professor Russell want a license system for AI, similar to aviation or medicine. He said governments might act only after a very big disaster happens.
