Altman Defends OpenAI's Pentagon Deal Amid Backlash

OpenAI's agreement with the Pentagon will deploy its AI models within the Department of Defense's classified networks. CEO Sam Altman has stated that the deal includes more safeguards than any previous classified AI deployment, including the one previously held by rival Anthropic. The deal came shortly after the Trump administration directed the government to cease working with Anthropic, which was subsequently declared a "supply-chain risk". OpenAI has established three "red lines" that it says the Pentagon has agreed to: its technology cannot be used for mass domestic surveillance, to direct autonomous weapons systems, or for any high-stakes automated decisions. To enforce this, OpenAI will use a cloud-only deployment, which the company says prevents its models from being directly integrated into weapons systems. Cleared OpenAI personnel will also be involved in overseeing the implementation. The deal follows a public dispute between the Pentagon and AI company Anthropic. The Department of Defense had reportedly asked Anthropic to loosen some of its internal safety measures to permit "all lawful use" of its technology, which Anthropic refused, citing concerns about autonomous weapons and mass surveillance. This refusal led to the termination of their contract, which was worth up to $200 million. Sam Altman has publicly acknowledged that the deal with the Pentagon was "definitely rushed" and that the "optics don't look good." He has stated that the primary motivation for the swift agreement was to "de-escalate the situation" between the AI industry and the Department of Defense. Altman has expressed that a close partnership between the government and AI companies is "super important". The collaboration has sparked significant backlash, with some users on social media platforms like Reddit and X calling for a boycott of ChatGPT and sharing instructions on how to cancel subscriptions. The deal has also drawn criticism from within the tech industry itself, with hundreds of employees from Google and OpenAI signing a petition in support of Anthropic's initial refusal to compromise on its safety standards. The Pentagon's use of AI is not new, with initiatives like Project Maven dating back to 2017. Project Maven utilizes machine learning to analyze data from sources like drones and satellites to identify potential targets for human analysts. Companies including Google, Palantir, and Amazon Web Services have previously been involved with the project. Google, however, withdrew in 2018 after internal protests from employees. Legal experts have raised concerns that the "guardrails" in OpenAI's contract may not be as robust as they appear. They point out that terms like "mass surveillance" are loosely defined in U.S. law, which could create loopholes. There is also concern that what constitutes an "autonomous weapon" is based on Department of Defense directives that could change over time. Altman has stated that OpenAI would terminate the contract if the U.S. government breached the agreed-upon terms, though he added, "We don't expect that to happen." The deal is part of a broader push by the Pentagon to integrate advanced AI, with Elon Musk's xAI also reportedly having its Grok model cleared for classified use.