OpenAI Claims DeepSeek Stole Its Intellectual Property to Train Their AI Model

OpenAI has accused a Chinese company, allegedly DeepSeek, of using a machine learning technique called "distillation" to transfer knowledge from OpenAI's large AI model to a smaller one, violating OpenAI's terms of service. This accusation arose after OpenAI and Microsoft investigated and subsequently blocked accounts suspected of engaging in this activity. Despite the allegations, OpenAI declined to provide evidence linking DeepSeek to the misconduct. The situation highlights a perceived double standard, as OpenAI itself has faced numerous lawsuits over copyright infringement, including cases from The New York Times, Indian book publishers, and others.

The controversy underscores the ongoing debate over the use of copyrighted materials in AI training. OpenAI has previously acknowledged the necessity of using such materials for developing leading AI models, yet criticizes others for similar practices. Tools like Glaze and Nightshade have emerged to protect digital artwork from AI scraping and style mimicry, though OpenAI views these technologies as "abuse," indirectly confirming their effectiveness. The situation reflects broader tensions in the tech industry regarding data usage and intellectual property rights.

Key takeaways:

OpenAI has accused a Chinese company of using "distillation" to transfer knowledge from its AI model to a smaller one, violating OpenAI's terms of service.
OpenAI and Microsoft investigated and blocked accounts suspected of engaging in distillation, believed to belong to DeepSeek.
OpenAI faces numerous lawsuits over copyright infringement, highlighting a double standard in its stance on data usage for AI training.
Tools like Glaze and Nightshade are recommended to protect digital artwork from AI scraping and style mimicry, though OpenAI views these technologies as "abuse."

OpenAI Claims DeepSeek Stole Its Intellectual Property to Train Their AI Model

Key takeaways:

Comments (0)

Newsletter