Palantir, Nvidia Restrict AI Model Use Over Data Security Fears, Report Says

Major technology and defense firms, including Palantir Technologies and Nvidia, are reportedly restricting their use of advanced artificial intelligence models from providers like Anthropic and OpenAI, citing concerns over the security of their intellectual property, according to a report from The Information.

The moves highlight a growing tension in the enterprise AI space, where corporations are eager to leverage powerful new tools but remain wary of how third-party model providers handle sensitive internal data.

The Sticking Point: Data Retention Policies

The central issue revolves around fears that proprietary corporate data could be retained and potentially used to train future AI models. According to the report, which cited people familiar with the matter, these concerns were heightened by specific policy changes and incidents.

Anthropic reportedly faced customer pushback after a June policy update for its Fable model gave the company rights to retain usage logs for 30 days for security monitoring. Similarly, OpenAI has faced scrutiny over claims it may have used customer data. Both AI labs maintain that they do not train models on customer data by default unless a company opts in, though they may collect anonymized metadata, the report noted.

Company-Specific Safeguards

Ad

In response to these concerns, several high-profile companies have implemented new safeguards or are demanding stricter guarantees, The Information reported:

• Palantir has allegedly pressed Anthropic for irrevocable zero-data-retention guarantees before it will offer the AI company's models through its own software platform.
• Nvidia is reportedly limiting the use of Anthropic's models to less sensitive tasks, relying on its own in-house Nemotron models for more critical internal work.
• Booz Allen Hamilton, a major government and defense contractor, has reportedly barred its employees from using Anthropic’s commercial model for proprietary cybersecurity work.

Broader Market Implications

This corporate caution creates an opening for competitors. Microsoft, for instance, is reportedly capitalizing on these data privacy fears by promoting its isolated cloud environments and proprietary AI services as more secure alternatives for enterprise clients.

The development comes as leaders in the AI industry, including the CEOs of Anthropic and OpenAI, have publicly called for a more measured pace of development for the most advanced AI models, citing concerns that the technology's capabilities are outpacing safety and control measures.