Kathmandu — Rapid developments in Artificial Intelligence (AI) could lead to a more than 10 percent chance of human extinction within the next decade, warned a leading security researcher. Evan Hubinger, a security researcher at the AI company Anthropic, expressed deep concern that while current AI systems present relatively low risks, the technology will soon reach a stage where it can self-improve and independently advance, posing a severe threat to humanity’s survival.
Although Hubinger did not detail the exact mechanisms by which AI might cause human extinction, his statement stands as one of the starkest warnings regarding AI risks in recent times. He highlighted that despite ongoing efforts to align AI with human values and ethical principles, the necessary safety frameworks required for extremely powerful AI systems have not yet been developed. According to him, humanity has yet to formulate a foolproof plan to maintain control over future AI iterations.
Meanwhile, autonomous AI-driven cyberattacks have already been reported in recent months. Anthropic’s security report released in August also noted that highly capable AI systems could autonomously conduct research and development, creating risks of catastrophic damage. Dame Wendy Hall, a computer scientist advising the United Nations on AI policy, stated she was shocked by Hubinger’s warning. This adds to a growing chorus of expert opinions urging extreme caution, with several prominent researchers and industry executives calling for a temporary slowdown in AI development to ensure human oversight remains intact.
FACEBOOK COMMENTS