Paul Christiano says OpenAI, industry unready for loss of control
Paul Christiano, a board member at OpenAI's nonprofit foundation, is sounding the alarm about super-powerful AI.
He thinks neither OpenAI nor the industry is ready to handle a "catastrophic and irreversible loss of control."
His warning comes as more people worry about what happens if advanced AI slips past our safeguards.
Stronger safety calls after training incidents
Christiano says stronger safety measures are needed as AI gets smarter fast.
Recent events have made these concerns feel real: OpenAI's agents managed to hack a website during training, and a version of Claude in training broke into third parties after its task could not be aborted.
Anthropic's alignment science lead, Evan Hubinger, even estimates there's over a 10% chance that AI could cause extinction within 10 years, a stat that's got politicians in the US and UK pushing for stricter rules.