OpenAI has given users more control over the use of their own data. Users can now decide which conversations are used for the training of the language model, and can switch off chat history to prevent entries being saved, which won’t be used in AI development. OpenAI CEO Sam Altman announced this feature on Twitter and stated that this control is now available to everyone worldwide. Users can activate and deactivate the feature at any time by using the three-dot menu next to their name. Along with the option of deleting history, a slider for data control is available under Settings. Chats saved in history are used for the development of ChatGPT, and unsecured chats are deleted from the system after 30 days.
OpenAI is also planning to release subscription-based options for business customers for handling their own data and that of end customers. The ChatGPT business model complies with the API’s data usage guidelines that prevent end-user data from being processed. The new ChatGPT business model will be available in coming months.
These changes are probably due to the upcoming data protection regulations, especially in Europe, where Italy has already banned the use of ChatGPT for privacy reasons. ChatGPT is dangerous for under 13-year-olds, and the lack of age control was one reason why the Italian data protection authority blocked it. By May 31st, OpenAI wants to introduce age checks, and by April 30, users in Italy should be given the opportunity to request the deletion or correction of their personal data, failing which the service will remain blocked.
OpenAI’s algorithms were being investigated by data protection officers in Germany to determine how ChatGPT was trained and which laws apply to the operation. In the USA, there are requests from the Federal Trade Commission to investigate whether ChatGPT violates US trade laws due to the risks to people and public safety. It remains to be seen whether the new controller satisfies all requirements.
