OpenAI shares its contract language and 'red lines' in agreement with the Department of War

OpenAI has announced that its recent pact with the Department of War establishes stronger safety measures compared to an earlier agreement that led to Anthropic facing blacklisting. In a blog entry released on Saturday, OpenAI detailed specific clauses from its contract, emphasizing that its technology will not be utilized for mass domestic surveillance or to develop autonomous weapons or decision-making systems involving critical assessments like social credit scores. According to OpenAI, this new agreement incorporates extensive safety protocols that surpass those found in previous contracts for classified AI implementations, including those relevant to Anthropic. The company highlighted that they maintain complete control over their safety framework, utilize cloud deployment, ensure involvement from authorized OpenAI personnel, and have robust legal protections integrated into the agreement, alongside existing safeguards provided by U.S. law. This development follows Anthropic's recent designation as a supply chain risk after it refused to comply with the Department of War's stipulations regarding its AI model, Claude. In a statement, Anthropic maintained that it would not succumb to pressure regarding issues of surveillance or autonomous weaponry and expressed intentions to legally contest its risk designation. In its blog post, OpenAI contended that Anthropic's classification as a supply chain risk should be reconsidered, stating that they have clearly communicated their stance to the government. OpenAI's agreement was partly motivated by a desire to ease tensions between the Department of War and U.S. AI research labs, advocating for a collaborative future between governmental bodies and tech firms. OpenAI requested that similar contractual terms be extended to all AI laboratories, particularly urging the government to negotiate with Anthropic to alleviate the ongoing discord. The company expressed that the current situation is detrimental to the future of collaboration in AI. As part of the terms, OpenAI will retain oversight of the safety mechanisms it employs and has established strong protocols to prevent any misuse of its technology. If the government were to breach the contract, OpenAI indicated that it could terminate the agreement, although they do not anticipate such an occurrence. The ongoing conflict between the government and AI companies has raised significant ethical concerns, particularly regarding the military's use of AI technologies. OpenAI believes that by allowing government access to its models, it can provide national security professionals with enhanced tools to navigate emerging risks associated with AI. Meanwhile, Anthropic's Claude has gained popularity, recently topping app store charts, with numerous users, including high-profile celebrities, publicly announcing their intention to cancel subscriptions to ChatGPT following OpenAI's military agreement.