Key Points:
- OpenAI signed a deal to place AI on classified Pentagon networks.
- The contract bans domestic surveillance and the operation of autonomous weapons.
- The agreement arrived just after President Trump banned the rival company Anthropic.
- The Pentagon labeled Anthropic a supply-chain risk over strict safety rules.
OpenAI signed an agreement with the Pentagon to put its artificial intelligence on classified military networks. The company announced the deal on Saturday, claiming the contract includes strict safety rules. This move happened hours after President Donald Trump ordered the government to cut ties with rival startup Anthropic.
OpenAI leaders said their contract contains stronger protections than previous military agreements. They drew three strict red lines. The military cannot use OpenAI technology for mass domestic surveillance. They cannot use it to direct autonomous weapons. Finally, they cannot use the software for high-stakes automated decisions.
To enforce these limits, OpenAI uses a strict technical setup. The company keeps full control over its safety software. They deploy the models entirely through the cloud instead of putting them into military hardware. Only cleared OpenAI employees manage the system to monitor how the military uses the tools.
This situation contrasts sharply with the crisis facing Anthropic. On Friday, Trump demanded that federal agencies stop using Anthropic products. The Pentagon then labeled the startup a supply-chain risk. This severe designation effectively blacklists the company from lucrative government contracts.
The conflict erupted because Anthropic refused to drop its ethical limits regarding weapons and surveillance. The Pentagon wants maximum flexibility and dislikes it when tech creators restrict military operations. Anthropic called the government’s punishment illegal and promised to fight the decision in court.
Last year, the Defense Department awarded massive contracts to several major AI labs. Now, the government is showing it wants total control over the technology.
OpenAI warned it will terminate its contract if the military violates the terms. However, the company also defended its rival, telling the government that Anthropic does not pose a supply-chain risk.
