Overview
- Several Google DeepMind employees resigned this week in protest of a contract the company signed earlier this year that allows its AI to be used in classified Pentagon settings.
- Resigners including Sofia Liguori and Andreas Kirsch say Google’s internal governance and oversight are insufficient to ensure human control when its models are used in military operations.
- The contract lets Google’s technology be used for "any lawful government purpose" while adding written limits on domestic mass surveillance and on autonomous weapons without human oversight.
- Google says it follows its AI principles, reviews public-sector deals through compliance channels, and provides employees safe reporting routes and anti-retaliation policies.
- The dispute comes as the Pentagon seeks faster adoption of commercial AI and several major vendors have signed similar agreements, raising questions about transparency, worker voice, and future oversight of military AI uses.