BREAKING
SBI posts robust Q1 on strong loan growthShooting claims life of Udupi Congress leaderRetail, mutual funds power record DII surgeIndia rejects US criticism on FCRA Bill changesNews Updates 07.08.2026: UPI to remain free for consumers, small merchants, says Payments Council of IndiaBrent crude inch higher to $79.90/bblMumbai cop deployed outside Salman Khan’s house collapses, diesKolkata to witness major traffic curbs on August 10 for ‘Har Ghar Tiranga’ rallyAccused No. 14 seeks to turn approver; Darshan objectsCorporate bond mobilisation falls 8.4% in FY26: Sebi
Technology

Anthropic fixes Claude privacy issue after private AI chats appeared in Google Search

hellohimachalse@gmail.com 28 Jul 2026 28 Jul 2026

New Delhi: Anthropic has fixed an issue that briefly exposed some shared Claude chatbot conversations in Google Search, raising fresh concerns over AI privacy. The incident surfaced when users noticed that search results led to chats with said content, such as cryptocurrency wallet information and personal information.

A Fortune report said the issue was related to Claude’s “share” feature, which enables users to create a public link for a conversation. Anthropic states it didn’t intend to make chats searchable, but that shared links might be found by search engines or archived by third parties.

Anthropic says users control shared conversation links

In response, Anthropic told Fortune that it is not adding chat directories or sitemaps to its search engines. Users will be able to decide if they want to share conversations publicly or not, the company said.

“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 with search engines like Google,” an Anthropic spokesperson told Fortune.

How the issue surfaced

Reddit users discovered that some Google search results included links to Claude conversations that had been shared and lists of Claude artefacts. Since then, Anthropic has taken the indexed links from Google Search. But old URLs that have already been shared will still be usable by anyone who already knows the URL.

AI privacy concerns continue

This ransomware is a reminder of an issue that AI companies are facing, particularly when it comes to data privacy. Fortune reports that OpenAI’s previous bot Skype had similar issues, allowing them to capture almost 100,000 conversations with ChatGPT, and that Elon Musk’s Grok bot had a similar problem.

Conversations with chatbots may reveal extremely sensitive details like work-related matters, monetary concerns, legal details, and personal issues, making them prone to disclosing information, according to privacy experts. This event highlights the need for improved security measures around public sharing capabilities on AI platforms.

Related News