Skip to content

OpenAI won’t use customer data anymore to train models

  • by
  • 2 min read

Photo: Koshiro K/Shutterstock.com

Soon after launching the ChatGPT and Whisper APIs, OpenAI addressed user and developer criticism by changing its API developer policy to ensure that any data submitted through the API won’t be used for “service improvements” including model training by default.

The company is also implementing a 30-day data retention policy for its API users, with options for stricter retention policies based on the user’s requirements. Finally, the company is also simplifying its terms and data ownership to clearly communicate that users own the input and output of the models. 

According to TechCrunch, Greg Brockman, the president and chairman of OpenAI, emerging legal challenges around generative AI models as well as user feedback, had a part to play in this policy rewrite. Brockman went on to state that one of the company’s biggest focuses is to become “super friendly to developers” in addition to building a platform that “others are able to build businesses on top of”.

OpenAI itself says that users shouldn’t share sensitive information with ChatGPT as they aren’t able to delete specific prompts from user histories. The old data retention and processing policy wasn’t much appreciated by developers who claimed that it helped the company profit off of their data and posed privacy risks. 

The aforementioned changes aren’t everything that OpenAI is changing either. The company is also changing its pre-launch review system in favour of a mostly automated system as its monitoring capabilities have significantly improved since last year. 

The old system required developers to wait in a queue for their app idea to be approved in concept. This new system lets developers build their apps and later identifies and investigates any problematic apps by monitoring their traffic. This should significantly reduce the burden on OpenAI’s already stressed review team while also allowing it to approve developers and apps using its APIs in a much higher volume.

In the News: Olympic Esports Week will feature Chess and Just Dance

Yadullah Abidi

Yadullah is a Computer Science graduate who writes/edits/shoots/codes all things cybersecurity, gaming, and tech hardware. When he's not, he streams himself racing virtual cars. He's been writing and reporting on tech and cybersecurity with websites like Candid.Technology and MakeUseOf since 2018. You can contact him here: yadullahabidi@pm.me.

>
Exit mobile version