OpenAI Accuses Chinese Firms of Copying ChatGPT AI Models After DeepSeek’s Cost-Effective Breakthrough"

WASHINGTON — OpenAI has accused Chinese companies of attempting to replicate its advanced AI models, following a stir caused by Chinese startup DeepSeek’s new chatbot, which rivals US models at a fraction of the cost.

DeepSeek's impressive performance has raised suspicions that it reverse-engineered AI models like ChatGPT. OpenAI claims rivals use a method called "distillation," where smaller AI models learn from the behavior of larger, more advanced ones.

A spokesperson for OpenAI said, "We know companies, including those in China, are constantly trying to distill models from US AI leaders." OpenAI is working closely with US authorities to secure its technology from such actions.

David Sacks, AI czar under the new Trump administration, echoed the claim, suggesting DeepSeek “distilled” OpenAI's models. Despite facing similar accusations over intellectual property violations, OpenAI emphasized that this practice violates its terms of service.

However, some experts, like Cornell’s Lutz Finger, argue that the use of copyrighted material in training models is a broader issue, noting that knowledge is hard to protect once it’s out in the world.