May 6 news: OpenAI CEO Sam Altman said on Friday that they have not used the data of paying customers "for some time." To train its GPT and other artificial intelligence large language models.
Altman said, "Customers have made it clear that they do not want their data used for training, so we have changed our plans: we will not do that."
OpenAI's commercial customers include Companies such as Microsoft, Salesforce and Snapchat call OpenAI’s software through application programming interfaces (APIs). Records from the Internet Archive’s Wayback Machine show that OpenAI quietly updated its terms of service on March 1 this year. Altman stressed: "We haven't used any API data for training in a while."
However, OpenAI's latest privacy and data protection measures only apply to customers using the company's API services. The company’s updated terms of use state: “We may use content from the Service other than content from our API.” This means that text messages that OpenAI users enter into ChatGPT may be used. According to reports, Amazon recently warned employees not to share confidential information with ChatGPT, fearing that the information might appear in feedback answers.
While OpenAI adjusts data protection policies, industries are grappling with the trend that large language models may replace human-created content.
For example, the Writers Guild of America has been trying to restrict the use of ChatGPT in the script creation or adaptation process. The strike began on Tuesday after negotiations between the association and the studios broke down.
Executives are also concerned about the impact of ChatGPT and similar projects on intellectual property. Entertainment tycoon and IAC chairman Barry Diller said media companies could take the issue to court and sue AI companies over the use of creative content.
The above is the detailed content of API access is safer, OpenAI says it will not use customer data for training. For more information, please follow other related articles on the PHP Chinese website!