OpenAI claims DeepSeek trained R1 using its data

A Href=” by Cyber Daily | Read the full article | Published January 30, 2025

OpenAI has raised concerns about a new AI model called R1 developed by a competitor named DeepSeek. The company claims that DeepSeek may have used its data without permission to train this new model, which has already made a significant impact on the market. In fact, the introduction of R1 led to a nearly $1 trillion drop in NVIDIA’s market value. Unlike OpenAI’s ChatGPT, DeepSeek’s R1 is open source, allowing other companies to build their own AI tools based on it.

The process that DeepSeek allegedly used is known as “distillation,” where knowledge from a larger AI model is transferred to a smaller one. While this technique is common in the industry, OpenAI’s terms of service prohibit competitors from using it to replicate their models. OpenAI has stated that they are actively monitoring the situation and are working with the U.S. government to protect their technology from potential misuse.

Microsoft’s security team is also looking into whether DeepSeek accessed OpenAI’s API to gather data for training R1. As the competition in the AI field heats up, the implications of these developments could have far-reaching effects on the industry and its regulations.

Read More