OpenAI hasn’t trained its AI large-language models akin to GPT with paying customer data “for some time,” CEO Sam Altman told CNBC on Friday.
“Customers clearly want us not to coach on their data, so we have modified our plans: We won’t try this,” Altman told CNBC’s Andrew Ross Sorkin.
OpenAI’s terms of service were quietly updated March 1, records from the Web Archive’s Wayback Machine show. “We do not train on any API data in any respect, we’ve not for some time,” Altman told CNBC. APIs, or application programming interfaces, are frameworks that allow customers to plug directly into OpenAI’s software.
OpenAI’s business customers, which include Microsoft, Salesforce and Snapchat, usually tend to make the most of OpenAI’s API capabilities.
But OpenAI’s latest privacy and data protection extends only to customers who use the corporate’s API services. “We may use Content from Services aside from our API,” the corporate’s updated Terms of Use note. That would include, for instance, text that employees enter into the wildly popular chatbot ChatGPT. Amazon reportedly recently warned employees to not share confidential information with ChatGPT for fear that it would show up in answers.
The change comes as industries grapple with the prospect of large-language models replacing material that humans create.
The Writers Guild of America, for instance, began striking Tuesday after negotiations between the Guild and movie studios broke down. The Guild had been pushing for limitations on using OpenAI’s ChatGPT for script generating or rewriting.
Executives are equally concerned concerning the impact of ChatGPT and similar programs on their mental property. Entertainment mogul and IAC chairman Barry Diller has suggested that media corporations could take their issues to the courts and potentially sue AI corporations over using the creative content.
