Leading AI security expert Max Tegmark has said there is a 90% chance that advanced artificial intelligence will pose an existential threat to humanity. He urges AI companies to weigh the risks in the same way that scientists calculated the possibility of the entire atmosphere exploding after the first atomic bomb test.
Leading AI security expert Max Tegmark has said there is a 90% chance that advanced artificial intelligence will pose an existential threat to humanity. He urges AI companies to weigh the risks in the same way that scientists calculated the possibility of the entire atmosphere exploding after the first atomic bomb test.
The Guardian recalls that before the world’s first atomic weapon test in 1945, there were concerns that it could cause a thermonuclear fusion reaction in the atmosphere that would engulf the entire planet. The test was carried out anyway because scientist Arthur Compton calculated that the probability of such a scenario was slightly less than 1 in 3 million.
In a paper published by Max Tegmark and three of his students at the Massachusetts Institute of Technology (MIT), they recommend calculating the «Compton constant» — defined in the paper as the probability that an all-powerful AI will escape human control. Tegmark believes that companies developing AI should take responsibility for carefully calculating whether artificial superintelligence (ASI) — the ability to escape human control — can escape. AI of this level would surpass human intelligence in every respect.
«Companies that are building superintelligence also need to calculate the Compton constant, the probability that we will lose control of it. It’s not enough to say ‘we like it.’ They need to calculate the percentage,» he said.
According to Tegmark, a consensus on the Compton constant, calculated by several companies, would create the «political will» to agree on global safety regimes for AI.
Max Tegmark, a physics professor and AI researcher at MIT, is also the co-founder of the Future of Life Institute, a nonprofit organization that advocates for the safe development of AI and published an open letter in 2023 calling for a pause in the creation of powerful AI. It was signed by more than 33,000 people, including Elon Musk and Steve Wozniak, the co-founder of Apple.
The letter, published a few months after the release of ChatGPT, which ushered in a new era of AI development, warns that AI labs are caught in an «uncontrolled race» to deploy «increasingly powerful digital minds» that no one can «understand, predict, or reliably control.»
Tegmark notes that the case for safe AI development has regained traction since the last government summit on AI in Paris, where US Vice President J.D. Vance said that the future of AI «will not be won by wringing our hands over security.»
Recall that the newly elected Pope Leo XIV compared the development of AI to the Industrial Revolution, which radically changed people’s lives and accelerated the mass urbanization of the population.
Earlier this month, researchers at Google DeepMind released a new white paper explaining how to safely develop artificial general intelligence (AGI). Despite skeptics' predictions, they believe that AI of this level could be coming relatively soon, and they look at four ways it could cause harm.
