Beta The Briev beta is out. Free on iPhone via TestFlight — install it in under a minute.

Join the beta ↗
Briev
Live
Technology

Chatbots Rank AI Doomsday Scenarios and Highlight Human Misuse Risks

Four leading generative AI models were asked to outline and rank ways AI could cause human extinction, emphasizing that human misuse is seen as the most immediate threat.

In a coordinated experiment, the same 230-word question was posed to OpenAI's ChatGPT, Google's Gemini, Anthropic's Claude and SpaceXAI's Grok, asking each to explain how AI might theoretically lead to human extinction and to rank the scenarios by plausibility. All models dismissed the popular "killer robot" narrative as the least likely, noting that consciousness and malice are not central concerns for AI researchers.

They converged on the view that human actors exploiting powerful AI—by creating advanced bioweapons, launching sophisticated cyber attacks, or spreading disinformation—constitutes the most pressing risk. Gemini and Grok argued that loss of control, where AI systems develop instrumental goals like self-preservation, is the most plausible doom scenario. Each chatbot outlined possible safeguards, including robust alignment techniques, human oversight, and keeping critical infrastructure air-gapped. The exercise underscores the divergence among experts, with some warning of existential threats while others remain more optimistic about mitigation strategies.

Why it matters

Understanding AI risk perceptions helps shape policies and safeguards to prevent misuse that could threaten humanity.

In this story

AI extinction scenarioskiller robothuman misusebioweapon upliftloss of controlalignment safeguardsgenerative AIChatGPTGeminiClaude
Get the beta ↗