ChatGPT will now allow its users to delete their history after 30 days because of privacy concerns.
When chat history is turned off, conversations had on the platform won’t be used to train the AI bot.
The conversations will be retained for 30 days before being permanently deleted, though OpenAI’s ChatGPT may monitor conversations for abuse. Users can do this by logging into their account and clicking the three buttons next to their email. Then, they click settings and another window will open. Here, users can select chat history and training settings to switch off a toggle labeled “enable chat history”.
A statement from ChatGPT said: “We hope this provides an easier way to manage your data than our existing opt-out process.”
Businesses and enterprises will soon be able to access a ChatGPT Business subscription that will follow the platform’s API data usage policies, so their users’ data won’t unintentionally be used to train the AI model.
The AI chatbot will also release an option to allow users to export their ChatGPT data to see what information it stores.
The new feature were introduced after concerns were raised about a reported data breach, leading Italy to ban the Microsoft-backed chatbot. The Italian watchdog said it would investigate whether the platform complied with GDPR, following an alleged data breach.
The data breach allowed conversation histories and payment information to leak, with OpenAI CEO Sam Altman saying he regretted the leak, and that the company had fixed the problem.
The watchdog said it could not legally justify “the mass collection and storage of personal data for the purpose of ‘training’ the algorithms underlying the operation of the platform".
It also said that ChatGPT did not have a filter to check the age of its users, which “exposes minors to absolutely unsuitable answers compared to … their degree of development and self-awareness”.
However, OpenAI told the BBC it had complied with privacy laws.
