OpenAI has unveiled a new agreement with the Pentagon that enhances safety measures for military AI applications, marking a significant shift from past contracts. This announcement follows President Trump’s decision to cut ties with AI competitor Anthropic due to technology safety issues, with the Pentagon labeling Anthropic a supply-chain threat. The new accord emphasizes robust protections for classified use, prohibiting OpenAI’s technology from being employed in extensive domestic surveillance, autonomous weapons systems, or critical automated decision-making. OpenAI claims their partnership has more stringent guardrails than previous agreements, highlighting a multi-layered safety approach and full discretion over their safety stack. The contract reinforces the Pentagon’s need for flexibility in AI deployment, while OpenAI warns that any violations could lead to contract termination. This development comes as the Pentagon continues to invest in AI partnerships, seeking to balance operational capabilities with safety standards in military applications.
Source link
Share
Read more