Sam Altman says OpenAI shares Anthropic's red lines in Pentagon fight (AI safeguards)

OpenAI CEO Sam Altman has publicly endorsed Anthropic's core ethical positions in its ongoing standoff with the Pentagon, creating a united industry front on AI safety for military use. Altman stated that OpenAI shares Anthropic's 'red lines' against developing AI for mass surveillance and lethal autonomous weapons systems. This alignment between two of the world's leading AI labs marks a significant moment in the industry's attempt to self-regulate its most powerful technologies, particularly as governments globally seek to integrate advanced AI into defense and intelligence operations. The support comes as both companies navigate complex negotiations with the U.S. Department of Defense (DOD), balancing commercial opportunity with their stated safety principles.

In parallel, OpenAI is working to finalize its own agreement with the DOD, proposing a model centered on technical safeguards over purely contractual ones. A key proposal involves 'cloud-only deployment' of its AI systems, which would prevent the military from running models on its own secure, air-gapped servers and instead keep them on OpenAI's infrastructure where usage can be monitored and controlled. This technical approach aims to create enforceable guardrails against prohibited uses like autonomous targeting. The strategy highlights a growing industry belief that, for high-stakes applications, software-based controls may be more reliable than legal agreements alone. The outcome of these negotiations will set a critical precedent for how commercial AI labs engage with government defense contracts worldwide.