Archive — US — techcrunch
Google expands Pentagon’s access to its AI after Anthropic’s refusal
Google has provided the Pentagon access to its AI technology for classified networks, allowing unrestricted use, after Anthropic refused similar terms. This follows Anthropic's resistance against the Trump administration over ethical boundaries for AI use, resulting in a lawsuit with the DoD. Despite internal opposition from employees, Google’s deal mirrors OpenAI's, with ambiguous non-use provisions for surveillance and weapons.
Key points
- Google grants the Pentagon access to its AI technology.
- The deal allows for all lawful uses, without the restrictions sought by Anthropic.
- Anthropic's refusal led to a supply-chain risk designation by the DoD.
- Anthropic is currently involved in a lawsuit with the DoD regarding its designation.
- Google's agreement includes vague language regarding the non-use of AI for surveillance or weapons.
- 940 Google employees signed a letter opposing sales to the DoD without ethical guarantees.
Why it matters
- This event reflects ongoing tensions between tech companies and government agencies over ethical AI use.
- The trend of tech companies supplying AI to defense sectors has grown, especially post-2020.
- Historical context includes debates on AI ethics stemming from military use since the early 21st century.
- The increasing concern about surveillance and autonomous weapons is central to current AI governance discussions.
Google's decision to partner with the Pentagon highlights the growing divide on ethical AI use between tech companies and government demands, responding to both commercial interests and employee concerns.