ChatGPT Conversations Leak Into Google Search Console
Personal ChatGPT conversations have been leaking into Google Search Console, raising privacy concerns. The issue, reported by Ars Technica, involves sensitive queries appearing in the tool used by webmasters to monitor search traffic. Jason Packer of Quantable discovered the problem, noting 200+ queries, including intimate and business-related chats, were exposed without user consent.
The leak was first flagged by Jason Packer, owner of analytics consulting firm Quantable, in a detailed blog post. He reported seeing over 200 queries, including some quite unusual ones, appearing in Google Search Console. These queries, often related to personal or business issues, were generated by users who expected their chats would remain private. Packer noted that users did not have any option to prevent the exposure of their chats, raising significant privacy concerns.
Packer’s blog post also linked the issue to earlier reporting by The Information, which cited sources claiming that OpenAI was scraping Google search results to power ChatGPT responses. These sources suggested that OpenAI was relying on Google to answer prompts seeking information about current events, such as news or sports. Packer questioned whether OpenAI had overlooked the privacy implications of this practice or if the company simply did not care. He pointed out that some of these searches relied on Google, mistakenly sending whatever the user typed in the prompt box to the Search Console.
OpenAI has acknowledged the issue but has declined to provide precise numbers on how many users have been affected. The company has stated only a small number of queries were leaked, but these figures have not been made public. This lack of transparency raises further concerns, given the massive user base of ChatGPT, which numbers in the hundreds of millions.
Perhaps most troubling to some users, whose identities are not directly linked in chats unless their prompts include identifying information, there does not seem to be any way to remove the leaked chats from Google Search Console. This means that the exposed data remains accessible, potentially exposing private information to unintended audiences.
The incident has sparked broader discussions about data privacy and the ethical responsibilities of AI developers. With the increasing use of generative AI models like ChatGPT, ensuring user privacy and data security has become a critical challenge. The incident serves as a cautionary tale for companies handling sensitive user data, highlighting the need for robust privacy protections and user control over their information.