AI Industry Grapples with Data Privacy Concerns as Companies Restrict Model Use
The AI industry is facing growing concerns over data privacy as some large companies restrict their employees' use of advanced models. Anthropic and OpenAI are at the center of the controversy, with several firms demanding assurances that their proprietary data will not be used to train AI models.
Anthropic's change to its flagship Fable model in June has raised red flags among customers. The company can now retain customer data, which some interpret as a means to ensure the model is not misused. However, others worry that sensitive business data may be caught up in the sweep.
Nvidia and Palantir have also restricted their employees' use of advanced models over privacy concerns. Nvidia has opted for its own in-house AI solution for tasks deemed too sensitive for Anthropic's Fable, while C Spire has agreements with both OpenAI and Anthropic that prevent them from using the company's data to train models.
Microsoft is trying to capitalize on these concerns by offering secure AI platforms. However, this approach comes at a cost, which may be a barrier for some companies. The lack of trust in AI companies has already led one U.S. utility company to cancel its plans to test Anthropic's Fable.