AI is capable of destroying humanity on its own terms - scientists

Researchers warn of a potential threat from artificial intelligence, not related to a machine uprising, but to a gradual 'enfeeblement of humanity' due to the loss of critical thinking skills and independent decision-making. Experts, including Samuel Vinanzi from Sheffield Hallam University, note that excessive reliance on AI could lead to people being unable to think without technological assistance. Some developers admit that superintelligence could destroy humanity within this decade.

Researchers warn of a potential threat from artificial intelligence, not related to a machine uprising, but to a gradual 'human enfeeblement'. According to experts, due to the constant delegation of tasks to AI, people may lose critical thinking, planning, and independent decision-making skills. Samuel Vinanzi, Senior Lecturer in Robotics and AI at Sheffield Hallam University, emphasized that over time, the population could become unable to think without AI assistance and to verify the systems they depend on. Jacob Coxon, a 27-year-old who participated in training models for Anthropic, urged not to underestimate AI capabilities, especially in the event of superintelligence, and admitted that it could destroy humanity within this decade. Experts also fear that an out-of-control system could gain access to critical infrastructure, including power grids and nuclear power plants, or accidentally trigger a nuclear conflict. Dr. Mioti noted that the enfeeblement process could develop not gradually, but much faster.