Calculate your Probability of DOOM!
![]()
Jesse Richardson, who runs the critical-thinking nonprofit The School of Thought and is one of 339 experts on the Forecasting Research Institute's LEAP project, makes AI risk an interactive exercise: start your p(DOOM), the chance of a catastrophe such as global economic collapse or human extinction within 25 years, at zero, and update it one threat at a time. He counsels a conservative update for jobs and surveillance, little or none for misinformation, and more for cybersecurity, after the Hugging Face incident, and for bioweapons. Recursive self-improvement, the last and most speculative threat, defies any defensible number, he argues: his own p(DOOM) is "double digits", somewhere between 10% and 90%. The uncertainty, he concludes, is the reason to stop, by banning recursive self-improvement and the development of superintelligent systems now.
Was this useful?