The chief executive of artificial intelligence company Anthropic has publicly accused rival OpenAI of dishonesty regarding its recent agreement with the U.S. Department of Defense. The allegation, reported by Bloomberg, centers on the messaging surrounding OpenAI’s decision to engage in military work, a domain Anthropic itself exited due to safety concerns.
Dispute Over Defense Contracts and AI ethics
According to the report, Anthropic CEO Dario Amodei characterized statements from OpenAI about its military contract as “straight up lies.” The conflict stems from a fundamental divergence in how the two leading AI firms approach work for defense and intelligence agencies. Anthropic previously terminated a contract with the Pentagon, citing internal disagreements over AI safety protocols and ethical applications.
Following Anthropic’s withdrawal, OpenAI moved to secure a partnership with the Department of Defense. The specific nature of this partnership has not been fully detailed in public disclosures. OpenAI has stated its work with the military is limited to non-offensive purposes, such as cybersecurity and veteran healthcare.
Background of Diverging Policies
The incident highlights a growing schism within the AI industry on the issue of military collaboration. Anthropic, a firm founded by former OpenAI researchers, has established a formal set of safety policies, known as its Constitutional AI framework. These guidelines explicitly prohibit the company from developing AI for use in warfare, surveillance, or other applications that could cause harm.
OpenAI, which initially had a ban on military use, revised its policies in early 2024. The update removed the blanket prohibition, allowing for work with certain government agencies, including the military, provided the applications align with the company’s broader mission and safety standards. This policy shift paved the way for the current Pentagon agreement.
Industry Reactions and Implications
Amodei’s sharp criticism, delivered at an internal company meeting, signals escalating tensions between the two competitors. The public airing of this dispute brings renewed scrutiny to the ethical boundaries of AI development and commercial partnerships. Industry observers note that the competition for government contracts, talent, and technological leadership is intensifying.
Neither OpenAI nor the Department of Defense has issued a formal response to Amodei’s specific allegation of dishonesty. OpenAI has previously defended its policy change as a necessary step to work on projects it believes will enhance national security in a responsible manner.
Looking Ahead
The public accusation is likely to fuel ongoing debates in Washington D.C. and within the tech community regarding appropriate oversight for advanced AI systems. Congressional committees examining AI regulation may seek further clarification from both companies on their internal safety reviews and contract stipulations. The competitive dynamic between Anthropic and OpenAI is expected to continue influencing their respective approaches to governance and commercial strategy, particularly concerning high-stakes government partnerships.
Source: Bloomberg
