How hundreds of Claude AI chats appeared in Google Search
What's the story
Anthropic's AI chatbot, Claude, has been embroiled in a privacy controversy after hundreds of user conversations were found on Google Search and other search engines. The issue raised alarms over user privacy as shared chat links containing personal and work-related information appeared in search results. Now, Google has clarified that it did not expose these conversations on its own accord.
Information exposure
Claude conversations were found on Google and other search engines
A BBC report revealed that Reddit users found over 200 Claude conversations through site-specific searches on Google. The same was seen on other search engines like Bing, Brave, and DuckDuckGo.
The conversations reportedly covered a wide range of topics from resumes with names, contact details, and employment history to unpublished work documents, healthcare-related research, and personal discussions.
Company response
Anthropic's response to the privacy concerns
In the wake of the privacy concerns, Anthropic clarified that users have full control over whether they share a conversation.
The company said that shared conversation links are "not guessable or discoverable unless people choose to share them themselves."
However, Anthropic admitted that once a conversation is publicly shared, it could be archived or indexed by third-party services like search engines just like any other public webpage.
Indexing explained
Google clarifies it does not decide which webpages are public
Google has clarified that it does not decide which webpages are made public. Instead, it is up to the website owners to decide if their pages can be crawled and indexed.
"We give site owners clear controls to decide whether pages can be crawled or indexed, and we always respect those directives," a Google spokesperson told BBC.
Content removal
Conversations removed from Google Search
Following the reports of indexed conversations, they were removed from Google Search. This suggests that Anthropic might have updated its website settings to prevent further indexing.
Similar controls can be implemented using standard search engine directives to block crawling or remove pages from search results.
Past occurrences
Similar incidents with ChatGPT, Grok chatbot last year
This incident isn't the first of its kind. Last year, OpenAI had to restrict the visibility of shared ChatGPT conversations after some were found through Google.
Elon Musk's Grok chatbot also faced a similar issue with a large number of shared conversations showing up in online search results.
The latest incident highlights the need for better understanding and management of public sharing features, especially when they involve potentially sensitive information like personal or confidential data.