Professor lost two years of “carefully structured academic work” in ChatGPT because of a single setting change
Marcel Bucher, a professor at the University of Cologne, lost two years of academic work because of one click in ChatGPT.
Marcel Bucher, a professor at the University of Cologne, lost two years of academic work because of one click in ChatGPT.
Bucher used ChatGPT Plus as an assistant for various tasks, including writing emails, structuring grant applications, reviewing publications, and analyzing student exam answers, writes PC Gamer.
The professor noted that he did not view ChatGPT as a source of accurate facts, understanding AI's tendency to make certain mistakes. Instead, he relied on the continuity and apparent stability of the workspace.
Last August, Bucher temporarily disabled the “data consent” option because he wanted to see if he would have access to all the model’s features if he didn’t provide OpenAI with his data.
“At that point, all my chats were permanently deleted and my project folders were purged—two years of carefully structured academic work gone,” says Bucher. “There was no warning. There was no option to cancel. Just a blank page.”
Bucher admits that he only kept partial copies of some of the conversations and materials, so the rest of his two years of academic work simply disappeared. The professor repeatedly contacted OpenAI support, but eventually received a response that the data was irretrievably lost and could not be recovered.
“This wasn’t a case of losing random notes or empty chats,” says Butcher. “It was an intellectual framework that was built over two years.”
Bucher argues that if a single click can irreversibly destroy years of work, then in his opinion and based on his own experience, ChatGPT cannot be considered completely safe for professional use. He noted that as a paid subscriber paying over $20 per month, he expected basic protections — including a warning about permanent deletion, a restore option (albeit limited in time), and backup or duplication of data.
Bucher concludes that OpenAI allegedly fulfilled its obligations regarding user privacy by deleting their data the second they requested it.
US court overturns ChatGPT's total data retention: OpenAI now only keeps logs of marked users
US court overturns ChatGPT's total data retention: OpenAI now only keeps logs of marked users
ChatGPT helped a teenager from Osaka carry out a cyberattack on a network of Internet cafes: how it was done
ChatGPT helped a teenager from Osaka carry out a cyberattack on a network of Internet cafes: how it was done
ChatGPT encouraged an American to stalk and harass women. The man hosted a podcast about it on Spotify
ChatGPT encouraged an American to stalk and harass women. The man hosted a podcast about it on Spotify
Read the country's main IT news in our Telegram
Read the country's main IT news in our Telegram
