Google has significantly revised its AI principles, sparking widespread discussion about its shift in stance regarding the use of AI in military and surveillance technologies. First established in 2018, these guidelines have long been lauded for their commitment to ethical AI development. However, in a recent update, the tech giant has removed critical pledges that once assured the public it would not “design or deploy” AI tools intended for weapons or surveillance applications.

Shift in Ethical Approach: What’s Changed?

The most notable change is the removal of a section previously labeled “applications we will not pursue,” which explicitly excluded AI for military use or technologies designed to harm individuals. The company has replaced it with a broader commitment under “responsible development and deployment,” which now emphasizes aligning AI with user goals, international law, and human rights, without the previously strict exclusion of certain applications. This adjustment suggests a more flexible approach, albeit with continued focus on ethical guidelines.

AI’s Evolving Role in Global Security

This change comes as AI increasingly becomes recognized as a “general-purpose technology,” prompting the company to rethink its earlier stance. Google executives, including DeepMind CEO Demis Hassabis, explained that AI’s potential to drive both economic growth and national security requires updated policies. They now emphasize the importance of democratic leadership in AI development and continued focus on AI’s benefits for global progress. They added that any AI development would be carefully assessed to weigh benefits against potential risks.

A Change in Response to Growing Pressures

The shift marks a departure from earlier concerns about the military use of AI. In 2018, Google had faced backlash for its involvement in Project Maven, a Pentagon initiative that would have used AI for drone surveillance, leading to the resignation of several employees. By 2021, however, the company had returned to pursuing military contracts, including a bid for the Pentagon’s cloud services, and reportedly worked with Israel’s Defense Ministry on AI projects.

The full details of this update can be found in the company’s blog post, which is available for reading here. Source