18% of AI Researchers Anticipated Extinction Scenario by 2024

In 2024, 18% of surveyed AI researchers estimated a probability of an extinction scenario due to AI developments. This statistic comes from a survey of 1,500 leading AI researchers, including Jacob Coxon from Anthropic and Daniel Selsam from OpenAI. Selsam described the situation as a "ticking time bomb," while a former DeepMind researcher warned that "AI could kill us all." These findings underscore a significant concern among experts regarding the risks associated with advanced AI systems. The sentiment reflects a growing anxiety within the AI community, which has been echoed in previous discussions about AI safety and governance. The increasing probability estimates suggest that apprehensions about AI risks are not only persistent but potentially escalating.