Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
Privacy

Nearly 100,000 ChatGPT Conversations Were Searchable on Google (404media.co) 13

An anonymous reader shares a report: A researcher has scraped nearly 100,000 conversations from ChatGPT that users had set to share publicly and Google then indexed, creating a snapshot of all the sorts of things people are using OpenAI's chatbot for, and inadvertently exposing. 404 Media's testing has found the dataset includes everything from the sensitive to the benign: alleged texts of non-disclosure agreements, discussions of confidential contracts, people trying to use ChatGPT to understand their relationship issues, and lots of people asking ChatGPT to write LinkedIn posts.

The news follows a July 30 Fast Company article which reported "thousands" of shared ChatGPT chats were appearing in Google search results. People have since dug through some of the chats indexed by Google. The around 100,000 conversation dataset provides a better sense of the scale of the problem, and highlights some of the potential privacy risks in using any sharing features of AI tools. OpenAI did not dispute the figure of around 100,000 indexed chats when contacted for comment.

Nearly 100,000 ChatGPT Conversations Were Searchable on Google

Comments Filter:
  • "set to share publicly" means "will be indexed and searchable". Is the issue that this is an obscure but default setting?

    • They're outraged that wealthy entities could access their customer's information without paying them for it.
    • by ceoyoyo ( 59147 )

      It is truly shocking that something users requested be shared publicly on the internet was shared publicly on the internet.

    • OpenAI originally stated that the links would not be searchable by search engines. The share button was supposed to give you a link that you could give specifically to whomever you wanted to share with, with at least a needle-in-a-haystack's level of privacy.

    • by allo ( 1728082 )

      It was a checkbox in the "share via link" dialog that was not checked by default. So not obscure and not a default. In the end you first need to click share, then you need to click the checkbox and then you need to click create link.

      I don't think there is a gallery of public chats, so you also needed to share the link in some place where search engines find it before it got indexed.

    • Is the issue that this is an obscure but default setting?

      That, and:
      - the fact that the general public doesn't understand the implications of having stuff left on "shareable"
      - in general settings that affect privacy should be "opt-in", not "opt-out", because not every single person online will be attentive to the thousand of tiny settings hidden behind dark patterns that needs to be flipped so their privacy isn't utterly violated.

      The UE could step in for the point 2 above.
      Reporting on such unintentional leaks could help raise awareness of point 1 above.

  • by Rosco P. Coltrane ( 209368 ) on Tuesday August 05, 2025 @11:47AM (#65567742)

    you're crazy.

    Remember this: all AI knows comes from stolen data. Why would you interactions with it be treated any differently?

  • by Mr. Dollar Ton ( 5495648 ) on Tuesday August 05, 2025 @11:52AM (#65567754)

    Nothing in an Eliza bot chat is worth reading.

  • So, people deliberately shared their chats and made them "discoverable," and now there's a ruckus about it? Why?!
    • dog days of summer. School is not back in session, people are getting the last of their summer vacay. All of Europe is on vacation right now. It's a very slow couple of weeks for news. If it bleeds - it leads - and this is as close to blood-in-the-water as you're going to get right now.
  • Inadvertent? I do not think you know what that word means.

    A researcher has scraped nearly 100,000 conversations from ChatGPT that users had set to share publicly and Google then indexed, creating a snapshot of all the sorts of things people are using OpenAI's chatbot for, and inadvertently exposing.

    USERS SET THEM TO SHARE PUBLICLY
    THAT IS NOT INADVERTENT
    IT IS A CHOICE

    TL;DR: GFY clickbait clowns

  • by allo ( 1728082 )

    And it was also a feature not a bug.
    People set a checkmark "Allow search engines to find this conversation" and search engines found the conversation. *surprised pikachu face*

HOLY MACRO!

Working...