OpenAI Accuses DeepSeek of Using Its AI Models Without Permission

Silicon Valley is in upheaval as OpenAI and Microsoft claim to have evidence that Chinese AI company DeepSeek used OpenAI’s API to train its own models—a potential violation of OpenAI’s terms of service.

How OpenAI Detected the Breach

Microsoft security researchers flagged suspicious activity in late 2024, noticing large-scale data extraction from OpenAI developer accounts linked to DeepSeek.

OpenAI believes DeepSeek used distillation, a process where smaller AI models learn from more advanced ones, to replicate its technology at a fraction of the cost.

The Irony of AI Data Scraping

OpenAI itself built its AI models by scraping massive amounts of internet data—often without explicit consent.

Now, it finds itself on the other side of the debate, accusing a competitor of unauthorized use of its technology.

Political and Security Implications

The allegations have caught the attention of U.S. officials, with former AI czar David Sacks warning of potential IP theft.

OpenAI has called for closer collaboration with the U.S. government to protect its AI models from foreign adversaries.

While OpenAI has yet to release detailed proof, this battle over AI ownership and security could have major implications for future AI regulation and international competition.