![\"A](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/03\/2cbc86e5dd2e327c5d1755211448f301.jpeg\")
A close-up of Sam Altman’s face in red tones, speaking into a microphone. The background is orange with a grid pattern, featuring several small black toy soldiers casting shadows. The image has a stylized, high-contrast, and graphic design appearance. <\/p>\n
OpenAI just handed one of its biggest rivals a massive PR victory, in a blunder that even CEO Sam Altman admitted had optics that \u201cdon\u2019t look good.\u201d<\/p>\n
On Friday, Altman announced<\/a> that OpenAI had reached a new agreement<\/a> with the Department of Defense over how its AI systems would be deployed across the military, an act that many saw as the company crossing the picket line. That\u2019s because Anthropic, a company founded by former OpenAI employees, had refused to give in to the Pentagon\u2019s demands that it give the military unrestricted use of its Claude AI, even as CEO Dario Amodei insisted that Anthropic\u2019s AI not be used for autonomous weaponry or the mass surveillance of US citizens.<\/p>\n Invest in Gold <\/p>\n Powered by Money.com – Yahoo may earn commission from the links above. <\/p>\n It was a move that could come at great cost for Anthropic. The Pentagon had vowed to ice the company out of contracts with the federal government by declaring it a \u201csupply chain risk,\u201d and even threatened to seize its tech.<\/p>\n But at least in the short term, it\u2019s OpenAI that\u2019s facing more blowback for its decision. Online, scores<\/a> of users<\/a> \u2014 ranging from your typical AI bro to, we kid you not, Katy Perry \u2014 are saying they\u2019re ditching ChatGPT in favor of Claude because of Altman\u2019s deal with the Pentagon. Indeed, Claude surged to the top of the App Store over the weekend, and as of Monday, still claims the number one spot above ChatGPT, which is currently in second place.<\/p>\n A recent thread in the r\/ChatGPT subreddit calling on users to to quit the AI chatbot quickly became one of the forum\u2019s most highly-upvoted posts of all time.<\/p>\n \u201cYou\u2019re now training a war machine,\u201d the thread reads<\/a>. \u201cLet\u2019s see proof of cancellation.\u201d<\/p>\n The fierce backlash is despite Altman claiming that the DoD agreement included the same restrictions that Anthropic had been seeking. But in the eyes of many of its users and critics alike, the fact that OpenAI had reached an agreement at all while Anthropic refused to bend the knee was a sign of its capitulation to a deeply unpopular administration. The ethics of a company that was founded on supposedly beneficent principles now allowing its AI systems to be deployed across the US military couldn\u2019t have faced a more immediate test when just hours after Altman announced the agreement on Friday, the US and Israel launched a series of deadly strikes in Iran that killed its leader Ruhollah Khomeini and hundreds of civilians<\/a>. (Reports suggest that the DoD used Claude to select targets in Iran<\/a>, meaning even Anthropic\u2019s principled stand may be yet more theater from the AI industry.)<\/p>\n Altman, meanwhile, has been in damage control. Following the deal\u2019s announcement, he hosted a rare AMA on X where he fielded questions about OpenAI\u2019s work with the \u201cDoW\u201d \u2014 referring to the \u201cDepartment of War,\u201d the Trump administration\u2019s preferred moniker for the DoD \u2014 and respondents didn\u2019t hold back.<\/p>\n \u201cHow did you go from \u2018a tool for the betterment of the human race\u2019 to \u2018let\u2019s work with the department of WAR\u2019?\u201d asked<\/a> one user. Another mocked<\/a> Altman by asking if he was happy that Claude overtook ChatGPT on the App Store. \u201cNo,\u201d Altman conceded.<\/p>\n One of the most pressing questions concerned what OpenAI would do if the DoD issued orders that violated the constitution, or sought to carry out mass domestic surveillance. Altman\u2019s line was that OpenAI would refuse any such orders, even if it meant imprisonment. (\u201cPlease come visit me in jail if necessary,\u201d he quipped<\/a>.)<\/p>\n But he also exhibited a blind faith that this would never be an issue by, more or less, extolling the virtues of the armed forces. Altman asserted<\/a> that the \u201cpeople in our military are far more committed to the constitution than an average person off the streets,\u201d and uncritically cited<\/a> a statement from a DoD official who vowed that it would never infringe on American\u2019s civil liberties or engage in \u201cunlawful\u201d surveillance.<\/p>\n Such pinky-promises from Trump administration figures were apparently enough to convince Altman that everything the military did or has ever done is entirely above board, to overlook the fact that the administration has leaned on cutting edge surveillance tech<\/a> to carry out mass deportations, and to memory-hole the name \u201cEdward Snowden.\u201d<\/p>\n \u201cI would also be terrified of a world where our government decided mass domestic surveillance was ok,\u201d Altman wrote<\/a> at one point. \u201cI don\u2019t know how I\u2019d come to work every day if that were the state of the country\/Constitution.\u201d<\/p>\n OpenAI users rightfully viewed Altman\u2019s feigned ignorance as an insult to their intelligence. \u201cYou cannot post the statements by an Administration that is known to lie and expect people to have trust or confidence in [you or your company],\u201d one fumed<\/a>.<\/p>\n At the end of the day, even Altman couldn\u2019t deny how the PR disaster he had created for himself. The DoD deal, he admitted<\/a>, \u201cwas definitely rushed, and the optics don\u2019t look good.\u201d<\/p>\n