OpenAI accuses DeepSeek of AI theft; Wall Street reacts
The American company OpenAI has accused the Chinese startup DeepSeek of using its proprietary AI models. According to reports by the British newspaper "Financial Times" on Wednesday, these models were allegedly used to train the artificial intelligence developed by the Chinese company.
OpenAI has accused DeepSeek of infringing its intellectual property, stating that it found evidence of a method known as distillation.
This process was created by developers to "achieve better performance on smaller models by using the output of larger, more powerful models, allowing them to achieve similar results at lower costs for specific tasks," explained bitget.com
Distillation is a common practice in the AI industry, frequently used by startups.
The Financial Times highlighted that in this case, however, it may have been used by DeepSeek to create its own competitive model, which constitutes a breach of the terms for using the artificial intelligence models offered by OpenAI.
The terms of service state that users may not copy any of its solutions or use the output data to develop models competing with OpenAI, stressed the "FT".
The American company refused to provide further explanation or details regarding the allegations against DeepSeek. Meanwhile, the Chinese startup did not respond to the newspaper's request for comment.
Shock on Wall Street
NVIDIA's shares fell by 17% on Monday, reducing its market value by $589 billion. This was related to the development of new AI models by DeepSeek.
The Chinese startup allegedly developed these models using much cheaper technology than that offered by NVIDIA, which is a company providing hardware for many artificial intelligence companies, including OpenAI.
