What did artificial intelligence answer when asked about the probability of humanity's extinction within a decade?
An AI system called "Noah Predict" estimated the probability that AI technologies will cause the end of humanity within a decade at about 0.85%, or one chance in 118, according to an experiment conducted by The Independent in collaboration with the Noah Predict system. The system analyzed the question from 41 angles and conducted 108 searches, based on 52 million data points. In contrast, some researchers believe the probability may exceed 10%. The system identified three main risk scenarios: malicious human use, labor market disruption, and loss of human control.
An artificial intelligence system named "Noah Predict" estimated the probability that AI technologies will lead to the end of humanity within the next ten years at about 0.85%, equivalent to roughly one chance in 118. The experiment was conducted in collaboration with the British newspaper The Independent. The system relies on analyzing millions of data points and news articles to track trends and assess risks, and it analyzed the question from 41 different angles and conducted 108 searches to verify information, drawing on approximately 52 million data points before reaching its final estimate. Although the percentage is less than 1%, it does not mean the risk can be ignored. In contrast, a number of researchers and specialists warn of much higher probabilities, with some believing the chance that AI will cause the extinction of humanity within the next decade may exceed 10%. Computer scientist Geoffrey Hinton, winner of the Nobel Prize in Physics, said a 10% probability is "not an unlikely estimate." The system identified three main risk pathways: first, malicious human use of AI in cyberattacks or the development of biological weapons; second, disruption of the labor market and institutions due to excessive reliance on automated services; and third, loss of human control over systems that may view human intervention as an obstacle. The system proposed measures to mitigate risks, including restricting system permissions and preventing their access to sensitive infrastructure.