OpenAI makes new options available for how to deal with your own data. On the one hand, the chat history can be switched off, on the other hand, users can decide which conversations flow back into the training of the language model. If the history is switched off, no entries are automatically saved, they do not appear in the sidebar and are not used for the further development of artificial intelligence (AI).
Like OpenAI in a blog post shared by CEO Sam Altman on Twitter, this control is now available to everyone worldwide. It can be activated and deactivated in the settings at any time. To do this, you have to use the three-dot menu next to your name. In addition to the option of deleting the history, which has existed for a long time, there is now also a slider for data control under Settings. New chats are saved in the history and used for the development of ChatGPT, it says and: “Unsecured chats are deleted from our system after 30 days.”
ChatGPT settings.
OpenAI has to face further privacy issues
Another announcement is the work on a subscription for business customers, where there will be more options for handling data – both for your own data and that of end customers. This will comply with the API’s data usage guidelines, which means that end-user data will not be further processed by default. This ChatGPT business model should be available in the coming months.
The background to these changes is likely to be the data protection regulations that are in the air, especially in Europe. Italy has already banned the use of ChatGPT for privacy reasons. The reason is that the users do not have sufficient insight into the data processing and cannot consent to it. Another problem that the Italian data protection authority faces is the lack of age control – ChatGPT is dangerous for under 13-year-olds. OpenAI wants to introduce an age check by May 31st. By April 30, users in Italy must be given the opportunity to request the deletion or correction of their personal data. Otherwise the blocking of the service remains. It remains to be seen whether the newly introduced controller satisfies the requirement sufficiently.
There are also studies in the USA. The Federal Trade Commission (FTC) is being asked to investigate whether ChatGPT violates US trade law due to the risk to people and public safety. In order to find out how ChatGPT’s algorithms were trained, the data protection officers of the federal states have also opened administrative proceedings against OpenAI in Germany. The question at stake is which legal basis applies to the operation at all.
(emw)
