The Irony Is Total - OpenAI Accuses DeepSeek of Stealing Data

OpenAI accuses Chinese company DeepSeek of using their AI output to train its own model - but isn't that exactly how OpenAI built its own business? The irony is striking. Is this the start of a bigger debate about AI companies' right to collect data?

Share
The Irony Is Total - OpenAI Accuses DeepSeek of Stealing Data

OpenAI and Microsoft have launched an investigation into the Chinese AI startup DeepSeek, whose language model R1 has drawn significant attention. According to reports from Bloomberg and Financial Times, OpenAI suspects DeepSeek may have trained its model on OpenAI's output - a method that potentially violates OpenAI's terms of service.

According to Bloomberg, the investigation is focused on whether DeepSeek gained unauthorized access to OpenAI's technology. "Such activity could violate OpenAI's terms of service or indicate that the group acted to remove OpenAI's restrictions on how much data they could obtain," the paper writes, citing sources familiar with the matter.

"Substantial Evidence" of AI Distillation

David Sacks, venture capitalist and member of the Trump administration, has also commented on the situation and says there is "substantial evidence" that DeepSeek used a technique called AI distillation to extract knowledge from OpenAI's models.

"It's a technique where one AI model learns from another by asking millions of questions, much like a student learning from a teacher," Sacks explained in an interview with Fox News. "The AI can thereby mimic the reasoning of the larger model and essentially absorb its knowledge."

OpenAI's Accusations - A Mirror of Its Own Strategy

The most striking part of this story, though, is the irony. OpenAI has built its entire business on collecting enormous amounts of data from the internet - often without explicit consent and sometimes in violation of terms of service. OpenAI's defense in the ongoing lawsuits against them has never been that they didn't collect data, but rather that they have the right to do so.

Now, when another player is alleged to have used a similar strategy against OpenAI, sharp accusations are suddenly being leveled. The question is whether OpenAI will succeed in defending its own data collection while criticizing DeepSeek for doing the same thing - or whether this marks the start of a deeper debate about AI companies' rights and obligations when it comes to data use.