OpenAI introduces ChatGPT incognito mode

OpenAI has created a ChatGPT “incognito mode” for users that do not want their conversation history to be saved or used to train the large-language model (LLM) AI chatbot.

The company said users will have the ability to choose whether or not to disable the new feature in a setting control under their accounts.

If disabled, conversations normally saved in ChatGPT’s history sidebar will no longer appear when a user begins a chat conversation.

OpenAI said the incognito mode can be reset at any time, giving the user complete control over their conversation privacy.

“We hope this provides an easier way to manage your data than our existing opt-out process,” OpenAI stated in Tuesday’s announcement.

The company did remind users that even when the chat history is disabled, OpenAI will “continue to retain new conversations for 30 days and review them only when needed to monitor for abuse, before permanently deleting.”

Additionally, users will be able to export a copy of their data via an export file that will include all conversations and other relevant data sent to the user by email.

The San Francisco-based startup also announced Tuesday a new ChatGPT business subscription that will give professionals “more control over their data as well as enterprises seeking to manage their end users.”

By default, the ChatGPT Business service will not integrate user data into training the AI models.

OpenAI's chief technology officer Mira Murati said the company will “be moving more and more in this direction of prioritizing user privacy."

"It’s completely eyes-off and the models are super-aligned: they do the things that you want to do," Murati said.

The CTO said OpenAI has depended on user information to help make its software more reliable, but added that the company still has challenges.

Several European nations have been scrutinizing OpenAI’s data collection policy since the large-language learning model exploded on the AI tech scene last fall.

The Microsoft-funded company has access to hundreds of millions of users' data, which have been used to continuously teach and and improve its AI models.

A data leak exposing some ChatGPT user payment information and chat histories to other users in March raised concerns from privacy advocates worldwide.

Following the leak, Italy banned ChatGPT citing possible privacy violations; France and Spain followed soon after with their own data privacy investigations.

Italy's privacy watchdog gave OpenAI until April 30 to meet a set of demands to reinstate ChatGPT's use in the nation.

Murati said the new features were not a result of Italy’s crackdown, but a way to put users “in the driver's seat" of their own data.

She also stated that OpenAI is compliant with all European privacy laws to date.

The company plans to roll out ChatGPT Business in the coming months.