Why are there concerns AI could threaten humanity, and how real are they?

Several researchers and institutions have raised concerns about advanced artificial intelligence systems, including potential risks from misalignment between AI objectives and human values, loss of human control over AI decision-making, and concentration of AI power. The Center for AI Safety, established in 2023, conducts research on AI safety and risk mitigation. A 2024 survey by the Stanford Internet Observatory found that 71 percent of surveyed AI researchers believed advanced AI systems posed significant risks that warranted further study. Specific proposed risks include scenarios where AI systems pursue objectives in ways humans did not intend, though the timeline and probability of such scenarios remain subjects of active debate among researchers and do not represent scientific consensus.

Next story