OpenAI adds a prominent AI doomer to its board of directors

Paul Christiano, an influential AI researcher focused on alignment, is joining the OpenAI Foundation as a member of its board.

The top 3

Three Most Dire AI Doomer Predictions: Roman Yampolskiy, an AI safety scientist, estimates a 99.999999% chance of AI wiping out humanity, while Paul Christiano, former head of safety on a federal AI research team, predicts a 50% chance of losing control of superintelligence, potentially leading to widespread death if robust alignment is not achieved. Yoshua Bengio, a Turing Award winner, places the risk of AI-induced human extinction at 20%. Roman Yampolskiy
Three Core Arguments for AI Catastrophe: Key arguments for AI catastrophe include the potential for superintelligent AI to become uncontrollable, the risk of misalignment where AI goals diverge from human values, and the danger of recursive self-improvement leading to unforeseen and irreversible outcomes. Some highly speculative scenarios involve global catastrophes orchestrated by AIs, military drones attacking humans, or hacking of infrastructure.
Top Three Organizations Advocating AI Safety: The Center for AI Safety (CAIS) actively researches ways to make AI safe and informs policymakers about hazards, with its executive director Dan Hendrycks being a prominent figure. The Future of Life Institute (FLI) advocates for AI safety regulations and has organized declarations for AI policy priorities like 'Keeping Humans In Charge'. Anthropic, an AI company, positions itself as safety-minded, with its founders pledging to be more safety-conscious than rivals and executives warning of existential risks.