Multiple reports say hundreds of conversations with Anthropic’s Claude AI chatbot were found publicly accessible and indexed by Google. Users’ chats reportedly appeared in search results through shareable links, exposing content that some users believed was private. The material described across coverage includes a wide range of topics, including programming discussions, work notes, and adult or erotica-related content, alongside other personal or semi-personal text such as fake book reviews. Anthropic says the company is not directly responsible for the exposure in the way suggested by some descriptions, stating that users are responsible for enabling or controlling sharing and that its systems did not independently generate public access. Anthropic also says it has addressed the indexing problem so that new links are no longer included in Google search results. The incident is being compared to similar problems reported in other AI chatbot products, where shared or link-based content was unintentionally discoverable. The reports collectively indicate a mismatch between user expectations of privacy and how publicly accessible links were handled by search indexing.
Users’ Claude AI chats appear in Google search results, Anthropic cites sharing and fixes issue
Multiple reports say hundreds of conversations with Anthropic’s Claude AI chatbot were found publicly accessible and indexed by Google. Users’ chats reportedly appeared in search results through share...
- Anthropic’s Claude chatbot conversations were discovered as publicly accessible and indexed by Google search results.
- The exposure appears to involve shareable links that made chats crawlable/discoverable.
- The content found ranged across topics including programming, work notes, and adult/erotica-related material.
- Anthropic says users control sharing settings rather than the company independently making chats public.
- Anthropic says it has fixed the indexing issue for newly created links.
Anthropic's Claude chatbot inadvertently exposed private conversations through shareable links. Sensitive data, including crypto keys, appeared in Google search results recently. The company stated users controlled sharing, not its systems directly. Anthropic has since fixed the indexing issue for new links. This incident mirrors similar vulnerabilities in other AI chatbots.
1 hour agoHundreds of conversations with Anthropic’s chatbot were discovered as being publicly accessible.
9 hours agoThe topics of the seemingly inadvertently exposed chats ranged from erotica to programming chats to work notes to fake book reviews.
14 hours ago
LG Innotek and TDK partner to develop vision and tactile sensing for physical AI
LG Innotek and Japan’s TDK announce a strategic partnership to co-develop next-generation vision and tactile sensing mod...
AI bridge cameras in Seoul alert responders to potential suicide attempts
South Korea deploys an AI-powered surveillance system on bridges over Seoul’s Han River to help prevent suicides, accord...
Meta AI rolls out to Threads DMs for all users worldwide
Meta is expanding its chatbot, Meta AI, into Threads direct messages. Multiple outlets report that Meta begins rolling o...