04.08.2025 10:40:00
Дата публикации
Last week, journalist Chris Stokel-Walker drew attention to the fact that the popular search engine Google was providing public links to people's conversations with ChatGPT.
This happened if a chat participant shared a link to their conversation with another person, such as a friend or colleague at work, and at the same time left the box for public discovery. Such a chat with AI became public and was indexed by Google and other search engines.
The situation attracted the attention of privacy experts - after all, in such public chats, users could discuss their health problems or commercial details with business partners.
A few days later, the chief information security officer of OpenAI reported that the "share chat" function had been disabled. Moreover, he added that the company is working on removing already indexed content.
OpenAI assumed that users would find useful conversations in Google, but did not take into account the risks associated with the fact that someone could "accidentally" share personal correspondence.
You can check for public links and delete them yourself in "Settings". You need to go to the "Data controls" section and select the "Shared links" block.
This case highlights the need for a critical approach to communicating with chatbots: avoid sharing sensitive personal information, since its public dissemination can be used against you. Most AI platforms explicitly warn that the entered data can be stored, analyzed, and transferred to third parties for training AI models.
Therefore, users should be aware that any messages containing vulnerable personal information lose their confidentiality status, and the AI platform itself does not provide guarantees of compliance with information security standards.
(the text is translated automatically)
This happened if a chat participant shared a link to their conversation with another person, such as a friend or colleague at work, and at the same time left the box for public discovery. Such a chat with AI became public and was indexed by Google and other search engines.
The situation attracted the attention of privacy experts - after all, in such public chats, users could discuss their health problems or commercial details with business partners.
A few days later, the chief information security officer of OpenAI reported that the "share chat" function had been disabled. Moreover, he added that the company is working on removing already indexed content.
OpenAI assumed that users would find useful conversations in Google, but did not take into account the risks associated with the fact that someone could "accidentally" share personal correspondence.
You can check for public links and delete them yourself in "Settings". You need to go to the "Data controls" section and select the "Shared links" block.
This case highlights the need for a critical approach to communicating with chatbots: avoid sharing sensitive personal information, since its public dissemination can be used against you. Most AI platforms explicitly warn that the entered data can be stored, analyzed, and transferred to third parties for training AI models.
Therefore, users should be aware that any messages containing vulnerable personal information lose their confidentiality status, and the AI platform itself does not provide guarantees of compliance with information security standards.
(the text is translated automatically)
