OpenAI CEO Sam Altman is scrambling to contain damage from a US military deal that sparked user backlash. The CEO admitted the agreement announcement was "rushed" and "sloppy," with internal communications revealing the company is amending the contract to explicitly prohibit using ChatGPT for domestic surveillance of US persons. This follows criticism over OpenAI's partnership with the Department of War.

The controversy has triggered a ChatGPT user exodus, with uninstall rates surging 295% according to Sensor Tower data. Users are flocking to rival Claude, which saw 37-51% more app installs after Anthropic walked away from its own military deal. The exodus coincides with declining ChatGPT response quality and GPT-4o model retirement, further fueling customer dissatisfaction.

Altman vowed he would "rather go to jail than follow unconstitutional orders and called on the government to reverse its directive freezing out Anthropic. The CEO's damage control attempt comes as OpenAI faces mounting pressure to balance lucrative government partnerships with its ethical commitments, a tension that may reshape the entire AI industry's approach to defense contracts.