OpenAI is having a rough week, U.S. uninstalls of its ChatGPT mobile app have reportedly jumped by around 294% day-over-day, a dramatic spike compared to its usual single-digit churn.
At the same time, one of its biggest rivals, Anthropic, is enjoying the opposite effect. Its AI assistant Claude has surged up the charts, even climbing to the No. 1 spot on Apple’s App Store in the U.S. For an AI race that usually revolves around model upgrades and benchmarks, this time, you can arguably point to the U.S. government as the catalyst.
So, what exactly happened? The backlash appears to be tied to OpenAI’s newly announced partnership with the U.S. Department of War (the Pentagon). Before OpenAI stepped in, Anthropic had been in talks with the Pentagon but ultimately walked away. The company raised concerns about the potential use of AI for mass surveillance and fully autonomous weapons, saying it wanted stronger safeguards in place.
Damilare Odedina
Multiple reports claims that when those terms couldn’t be agreed upon, the deal fell through. But shortly after, the U.S. government announced plans to phase out Anthropic’s tools across federal agencies.
Almost immediately after, OpenAI swooped in. The company quickly secured a similar agreement to deploy its models in classified environments. CEO Sam Altman later admitted on X that the deal was “definitely rushed” and that “the optics don’t look good.” He claimed the move as an attempt to de-escalate tensions between the government and the AI industry, arguing that OpenAI’s contract includes clear red lines around mass domestic surveillance, autonomous weapons systems, and high-stakes automated decision-making.
But people aren't buying it. App review data shows a sharp spike in one-star ratings for ChatGPT, while five-star reviews declined over the same period. Meanwhile, Claude’s downloads have surged, in some estimates even surpassing ChatGPT’s daily U.S. installs for the first time.
Anthropic has also leaned into the moment, highlighting its safety-first stance and even guiding users on how to switch platforms and port over information from other AI tools.
The ethics of AI in national security have always been complicated, but this episode has made the debate far more visible to everyday users. Whether OpenAI’s gamble pays off or further fuels skepticism remains to be seen. For now, though, the App Store charts, and uninstall numbers, suggest that public perception can shift just as quickly as the AI landscape itself.
Damilare Odedina
