ChatGPT Chats May Appear in Google Search Results

Recently, a significant number of private ChatGPT conversations surfaced in Google search results, largely due to the chatbot’s “Share” feature, which has now been discontinued following a public outcry. This alarming trend highlights the challenges of privacy in the digital age.

Fast Company reported that users of ChatGPT may have unintentionally made their conversations public and searchable online. The investigation revealed almost 4,500 ChatGPT discussions indexed by Google, covering sensitive subjects such as mental health, relationships, and other personal matters. Fortunately, the identities of the users linked to these posts remained anonymous, sparing them from potential repercussions.

How did these conversations end up on the web?

Until recently, users of ChatGPT had the ability to share their chats with friends, family, or colleagues by making them public. This feature functioned similarly to the sharing options available on a Google Doc, where users could generate a public link to their discussion. Additionally, there was an option that allowed users to make the post “discoverable,” which, unbeknownst to many, meant it could be indexed by Google search results.

When users opted to create a shareable link for their chats, they would see a pop-up notification indicating: “A public link to your chat has been created.” Beneath this message, a checkbox labeled “Make this chat discoverable” was presented, along with a disclaimer in fine print that read: “Allows it to be shown in web searches.” This potential for misunderstanding led to many users unintentionally exposing their private chats.

By selecting this checkbox, users inadvertently permitted their conversations to be indexed by Google, allowing the company’s web crawlers to access and include these pages in search results, leading to unintended exposure of personal discussions.

Following the Fast Company report, OpenAI swiftly removed the controversial feature, with an executive describing it as a “short-lived experiment.” The decision aimed to safeguard users from further unintentional breaches of privacy.

OpenAI‘s Chief Information Security Officer, Dane Stuckey, clarified on X how the sharing feature was intended to function and where it ultimately failed in terms of user privacy and security.

This Tweet is currently unavailable. It might be loading or has been removed.

Despite the fact that ChatGPT users needed to actively choose to make their chats public, OpenAI determined that the risk of user error was too significant to maintain the feature. This highlights the fine line between user autonomy and ensuring digital safety.

As previously reported by Mashable, OpenAI is mandated to retain user conversations, even those that users have chosen to delete, due to a legal challenge from the New York Times. This legal requirement compels OpenAI to archive all interactions indefinitely, with exceptions for ChatGPT Enterprise and ChatGPT Edu customers.

Therefore, while ChatGPT users can enable a “Temporary Chat” feature akin to an incognito mode in a web browser, it is crucial to understand that chat data may still be stored and retained by the company.

Disclosure: Ziff Davis, the parent company of Mashable, filed a lawsuit against OpenAI in April, claiming that it infringed upon Ziff Davis copyrights while training and operating its AI systems.