Deepseek may have used chatgpt data for his AI model, Openai increased the alarm
Openai has launched an alarm of Chinese competitors, Deepseek, showing that the company may be “inappropriate” has used its data to develop its own artificial intelligence chatbot. This requirement was given after DeepSeek released a large and low -cost language model (LLM) with low cost on Monday, surprising the United States Technology industry. The launch has led to a significant decline in NVIDIA’s warehouse, a major American chip manufacturer.
The surprising appearance of Deepseek
Deepseek’s sudden participation in AI race with the new LLM has attracted the attention of the global market, causing concerns in the US technology community. Model performance, especially at a competitive price, has made comparisons with the top model of Openai, Chatgpt. However, Deepseek’s release has made many people wonder how the company manages to develop such an effective AI in such a short time.
Openai distillation requirements
To respond to the launch, an Openai spokesman revealed that the company is investigating whether Deepseek may have used a controversial technique called “distillation” to develop its model. or not. Distillation includes training of an LLM by using data created by another LLM, a process that is often banned according to the terms of service agreements but is often used in the industry.
We are aware and consider signs that Deepseek may have been distilled improperly from our models and will share information as we know more, the spokesperson said in an email sent to. NBC News. The spokesperson clarified that the company did not accuse Deepseek about any security violations, but asked questions about the legality of its data activities.
The concern of the industry and the legal validity
Although distillation is not uncommon in the AI industry, it asks moral questions about intellectual property and data ownership. Openai himself has faced legal challenges for allegations that they used not yet licensed content, including millions of articles from the New York Times, to train their models. .
Interestingly, the CEO of Openai Sam Altman congratulated Deepseek on X (formerly Twitter) just a few days before the controversy, praising the company for the impressive model, especially for an affordable price. .
The participation of the potential White House
As tensions increase, Openai has shown that the US government, especially the White House, can participate in the dispute. The company has pledged to continue working with the US government to protect the intellectual property and capabilities of US AI companies.
