OpenAI says it won’t spy on user data to train its AI models

In a significant policy shift, OpenAI, the maker of popular generative language models Dall-E and ChatGPT, has announced that it won’t be using data fed by its API customers to train its AI models. This comes in response to the concerns over the data privacy and ethical use of data by AI models.

The company has introduced two major changes to its policy- first OpenAI has decided to change its policy regarding the use of data submitted by users via its API. The company says it will not use this data to train its models or improve its services unless the user explicitly allows it. In case user wants their data to be used to improve the models, they can choose to opt-in to share their data. For the data that has been submitted before March 1, 2023, the company explains that it may have been used for improvements unless user had previously opted out of sharing data.

Next, OpenAI says it keeps API data for up to 30 days to monitor for abuse and misuse. Only authorised employees and third-party contractors can access this data to investigate suspected abuse. Enterprise customers can request not to have their data stored, but OpenAI will still check if any suspected abuse. If users upload files to fine-tune their models, the files will be kept until they delete them.

OpenAI has also removed its pre-launch review and claims it has made its terms of service and usage policies more developer friendly.

Altman while sharing the news on Twitter wrote, “we frequently don’t get it right the first time, but we really do listen and try to rapidly improve. thanks a lot to all of the developers who have given us feedback and patience. more improvements coming!”

While OpenAI has been at the forefront of AI technologies and has given some very popular language models to the world, the company’s reliance on user data for training its models has drawn criticism from privacy advocates and developers.

These policy changes apply to only API customers. It isn’t available for OpenAI’s Non-API consumer services like ChatGPT or DALL-E Labs.