Former Google DeepMind Researcher Resigns in Protest of Military AI Contract with Department of Defense

While the initial goal was to stop the provision of services to the DHS, the scope of the activity expanded as the U.S. Department of Defense pressured AI companies to enter contracts without restrictions on the use of lethal robots or mass surveillance.

The author called for action from individuals such as Jeff Dean, who pledged opposition to supporting lethal robots in 2018, and Stuart Russell, who has been active in the movement against autonomous weapons.

The author stated that while Dean signed an amicus brief supporting Anthropic, he took no action to block Google's own contracts.

Russell promised to issue a statement, but according to the author, that promise vanished.

The author drafted a 25-page proposal including contract clauses and monitoring mechanisms and submitted it to CEO Demis Hassabis.

This proposal, evaluated by experts in military and surveillance law, was ignored after being passed to senior policy staff.

Although management had stated they would not enter such contracts, the author's warnings were largely ignored.

Google ultimately signed a contract to hand over AI without restrictions on lethal robots or mass AI espionage.

Upon learning that the contractual constraints were weaker than those of OpenAI, the author decided to resign, stating they could no longer remain with the company in good conscience.

The essay describes the situation as a story of how powerful individuals and institutions failed to uphold AI ethics promises under pressure.