Google has signed an agreement with the Pentagon allowing its artificial intelligence models to be used in classified operations, The Information reported.

According to the outlet, the deal gives the US military access to Google’s AI technologies for any lawful government needs. Under the agreement, the company must also help configure safeguards and filters if requested by US authorities.

The contract states that the technologies must not be used to monitor US citizens or in autonomous weapons without proper human control and oversight. At the same time, Google will not be able to restrict the use of its AI models within lawful operations.

A Google representative told Reuters that the company supports both classified and unclassified projects with government agencies. The company views Pentagon access to its AI infrastructure as a responsible approach to national security.

Before the contract was signed, more than 600 Google employees appealed to CEO Sundar Pichai not to provide the US military with AI models for classified operations. The petition argued that artificial intelligence should serve society and not be used in inhumane or highly harmful ways.