Google has signalled a stronger push into national- security partnerships, addressing employee concerns over its expanding work with the US Department of Defence. The move reflects a broader shift in how the company is positioning its artificial intelligence (AI) capabilities.
During an internal townhall at Google DeepMind, employees raised questions about how collaborations with defence agencies and firms such as Boston Dynamics align with the company’s AI principles. Leadership responded by emphasising that all projects undergo a structured review process focused on safety, accountability, and privacy safeguards.
The company highlighted that its AI principles were updated in 2025. A previous restriction on the use of AI for weapons or surveillance was removed, marking a notable policy shift. The current framework focuses on evaluating whether the overall benefits of a project outweigh potential risks.
Executives indicated that more government collaborations are likely in the near future. The approach reflects growing engagement with national- security priorities, particularly as governments seek advanced AI capabilities for operational efficiency.
This stance was reinforced by leadership at DeepMind, which expressed confidence in balancing innovation with responsibility. The company maintained that working with elected governments is part of ensuring technology is used to enhance global safety.
The renewed engagement comes years after Google stepped away from a Pentagon project in 2018 following internal protests. Since then, it has gradually re-entered the space. Recently, the company secured a contract to deploy AI tools across non-classified defence systems, focusing on administrative tasks such as document summarisation and contract analysis, rather than combat use.
The development comes amid rising competition. Firms such as Amazon and Oracle are also pursuing defence deals, while Anthropic has taken a more restrictive stance on military applications, highlighting growing divides within the tech industry.
