Chinese AI company DeepSeek is facing accusations of using OpenAI’s data to train its own models, sparking an investigation by OpenAI and Microsoft.
DeepSeek, a Chinese AI firm known for developing advanced models at a low cost, is under scrutiny for allegedly using data from OpenAI to create its AI technology. OpenAI and Microsoft are investigating the company after Microsoft’s security team uncovered evidence of data being transferred to China through OpenAI’s developer account in late 2024.
It is believed that DeepSeek employed a technique known as “distillation,” where data from larger, more advanced models is used to train smaller ones, enabling the creation of competitive AI models at a fraction of the cost. OpenAI suspects that DeepSec utilized the output from its GPT models, in violation of its terms of service, although specific evidence has not yet been disclosed.
David Sachs, a former advisor to President Trump on AI, remarked that DeepSeek could face intellectual property theft charges based on the evidence gathered. OpenAI has expressed its concerns, stating that competitors, especially Chinese companies, are increasingly attempting to replicate top US AI models.
Liberty News’ international correspondent reports while no formal legal action has been taken, both OpenAI and Microsoft continue to investigate the matter, signaling a rising emphasis on safeguarding intellectual property in the AI industry. The situation highlights the growing tensions surrounding AI development, ethics, and data usage. As AI competition intensifies, such incidents of potential data theft could become more prevalent, prompting stricter regulations moving forward.
LND/BAIZID
