OpenAI, the San Francisco-based start-up behind the hit chatbot ChatGPT, has introduced an "incognito mode" that will not save users' conversation history or use it to improve artificial intelligence, the company announced on Tuesday.

This new feature comes amid growing scrutiny over how ChatGPT and other chatbots manage users' data, which is commonly used to train AI. Italy recently banned ChatGPT for possible privacy violations, while France and Spain have also begun probing the service.

OpenAI's chief technology officer, Mira Murati, shared with Reuters that the company is compliant with European privacy law and is working to assure regulators. The new features did not arise from Italy's ChatGPT ban, according to Murati, but from a months-long effort to prioritize user privacy and put them "in the driver's seat" regarding data collection.

The new product release allows users to switch off "Chat History & Training" in their settings and export their data. Nicholas Turley, OpenAI's product officer, likened this to an internet browser's incognito mode, with the company retaining conversations for 30 days to monitor for abuse before permanently deleting them.

In addition, OpenAI plans to launch a "ChatGPT Business" subscription in the coming months with additional data controls. This subscription will not use conversations for AI model training by default.

User information has helped OpenAI improve its software, reduce political bias, and tackle other issues, but the company still faces challenges. Microsoft, which has invested in OpenAI, already offers ChatGPT to businesses. Murati said that this service would appeal to the cloud provider's existing customers.