ChatGPT's creator calls out DeepSeek for copying its work

Recently, ChatGPT, one of the world's most popular chatbots, was overtaken by the Chinese app DeepSeek as the top downloaded app in the US App Store. Shortly after, OpenAI, the creator of ChatGPT, expressed concerns about a possible intellectual property breach.

Reportedly, OpenAI has found evidence indicating that Chinese AI startup DeepSeek may have secretly used data generated by OpenAI's technology to enhance its own AI models. If this turns out to be true, DeepSeek would be violating OpenAI's terms of service.

The issue revolves around a process known as "distillation," where developers take outputs from larger AI models and use them to train smaller ones. While this is a common technique in the AI world, OpenAI argues that DeepSeek took it too far by using this method to create a competing model.

David Sacks, the White House's AI and crypto czar, shared in a recent interview that he's come across evidence supporting the claim that DeepSeek did, in fact, use OpenAI's models to enhance its own.


OpenAI hasn't shared any additional details or specifics about the evidence it found. However, its terms of service clearly state that users aren't allowed to "copy" any of its services or use outputs to create models that compete with OpenAI.

It's reported that OpenAI, along with its partner Microsoft, looked into accounts linked to DeepSeek last year. These accounts were suspected of using OpenAI's API for distillation purposes, and as a result, their access was blocked for breaching the terms of service.

DeepSeek came into the spotlight earlier this month with the launch of R1, a reasoning model designed to break down requests into multiple steps. R1 is built on DeepSeek's main AI model, V3, which the company claims to have developed for just over $5.6 million – a small fraction of the massive investments that giants like OpenAI, Meta, and Google usually pour into their AI projects.

I think it's crucial to safeguard the most advanced models from attempts by rivals looking to steal them. That being said, OpenAI itself is facing accusations of copyright infringement. Content creators and newspapers are claiming the company trained its models using their articles and books without consent.