OpenAI and Anthropic rehearsing responses to potential AI catastrophe

By Splitfeed EditorialLast updated Oct 10, 2026, 00:40 UTC

On October 9, 2026, reporting from Futurism and Decrypt News revealed that OpenAI and Anthropic are both conducting internal exercises designed to prepare their organizations for how they would respond in the wake of a major AI-related disaster. The exercises, described as being carried out quietly and without public disclosure, include planning for scenarios involving public revolt or backlash following a catastrophic AI incident. On the same date, John Belton of Gabelli Funds noted in a CNBC appearance that revenue at both companies remains a closely watched metric, placing the safety planning alongside ongoing commercial pressure.

The significance of this story lies in what the exercises imply. That two of the most prominent AI developers feel it necessary to rehearse crisis responses suggests internal acknowledgment that a serious incident is at least plausible enough to plan for. The fact that these preparations are being conducted without public disclosure raises questions about transparency: whether internal crisis planning of this kind is a responsible precaution or a sign that the companies believe risks are higher than their public communications suggest. The gap between what these organizations say publicly about safety and what they prepare for privately is the central point of genuine dispute.

With no published reactions yet from the Pro-AI, Anti-AI, or Middle Ground camps, each position can only be anticipated based on how each typically frames stories of this kind. Supporters of AI development would most likely argue that crisis preparedness is standard corporate practice and a mark of responsible governance, not evidence of danger. Critics of AI would be expected to treat the exercises as an implicit admission that catastrophic outcomes are taken seriously behind closed doors, contradicting reassurances offered to the public and to regulators. A middle ground perspective would likely welcome the planning as a sensible risk management step while calling for greater transparency so that outside parties can evaluate whether the preparations are adequate.

The details most worth tracking in the coming weeks include whether either company voluntarily discloses the scope or findings of these exercises, how regulators respond if they become aware of them, and whether Belton's continued focus on revenue at both firms intersects with any shift in how aggressively they deploy products despite identified risks.