Nearly one in five AI researchers already expected an extinction scenario from AI back in 2024
What happened
A 2024 survey of over 1,500 leading AI researchers found the average estimated likelihood of AI causing human extinction at 18 percent. The figure sparked renewed attention after Anthropic researcher Jacob Coxon raised the issue in a high-profile tweet. OpenAI’s Daniel Selsam described the risk as a “ticking time bomb,” and a former DeepMind expert warned AI could potentially end humanity. These estimates reveal a growing sense of urgency about existential AI risks within the research community.
Why it matters
This level of risk assessment changes the conversation from speculative sci-fi to tangible concern among those building the technology. Operators, investors, and regulators need to factor in that nearly one in five top AI experts see extinction-level scenarios as plausible within this decade. This shifts incentives around safety investments, governance structures, and the pace of AI deployment. Companies that ignore these warnings might expose themselves to catastrophic systemic risk, while those acting on them can reduce existential downsides and build trust.
What to watch next
Watch how leading AI labs and governments respond with concrete safety protocols or regulatory moves. The growing alarm might pressure labs to slow some aspects of model capabilities or increase transparency around risk mitigation. Also follow funding shifts toward AI safety research, which could get prioritized over product pushes. The evolving public and regulatory debate on AI existential risk will shape operational constraints and investment focus in the near term.
AI Quick Briefs Editorial Desk