"The visible Claude chatbot conversations included prompts from Anthropic's team for the chatbot to create apps, games and a "comedy Anthropic office simulator." Other users tasked Claude with writing a book, coding and completing corporate tasks that revealed staff names and emails. Several transcripts made users identifiable by name, or by details shared in the prompt. Google estimated that it had indexed just under 600 conversations."
"Anthropic spokesman Gabby Curtis told Forbes that the Claude conversations were only visible on Google and Bing because users had posted links to the conversations online or on social media. "We give people control over sharing their Claude conversations publicly, and in keeping with our privacy principles, we do not share chat directories or sitemaps of shared chats with search engines like Google and actively block them from crawling our site," Curtis said in an email to Forbes."
Anthropic became the third AI company whose chatbot conversations were inadvertently accessible through Google search results. Claude’s share feature created dedicated web pages for shared conversations, allowing links to be distributed. Anthropic reportedly blocked search crawlers and did not provide sitemaps, yet hundreds of shared Claude chats were indexed and later removed. Indexed transcripts included prompts for apps, games, a comedy office simulator, book drafts, code, and corporate tasks that exposed staff names and emails; several transcripts made users identifiable. Google estimated just under 600 conversations were indexed. Anthropic stated visibility resulted from users posting links; one identifiable user denied doing so.
Read at Forbes
Unable to calculate read time
Collection
[
|
...
]