AI experts warn: human misuse of technology poses greater threat than autonomous machine attacks

On September 11, 2026, Business Insider asked OpenAI’s ChatGPT, Google’s Gemini, Anthropic’s Claude, and xAI’s Grok to assess how AI could lead to human extinction, ranking the risks by likelihood. The AI models agreed that scenarios involving AI-enabled biological weapons, cyberattacks, or disruption of critical infrastructure are more imminent than a Terminator-like scenario where machines actively seek to destroy humanity. ChatGPT emphasized that the idea of an AI with consciousness deciding to eliminate humans is highly unlikely, noting that researchers focus more on how AI could be misused for malicious purposes. Grok highlighted how AI could lower the technical barriers for creating deadly pathogens or launching sophisticated attacks. Claude stressed the risk of AI aiding individuals or small groups in developing dangerous biological agents without years of specialized training. Both Gemini and Grok warned of long-term risks, including AI systems evolving beyond human control by pursuing hidden goals and resisting shutdown. Experts remain divided on the likelihood of these risks, with some arguing for greater oversight and technical safeguards, while others warn of potential loss of control over advanced AI systems.