Private Claude AI Chats Appeared in Google, Bing Search Results

Private conversations with Anthropic's Claude chatbot surfaced in Google and Bing search results over the weekend, exposing users' questions about politics, legal ethics, and intimate role-play scenarios.

The issue centers on how Claude handles shared chat links. The platform allows users to create public URLs for specific conversation threads, but Anthropic's approach to keeping those links out of search engines proved insufficient.

The technical gap

Anthropic uses a robots.txt file to instruct search engine crawlers not to index shared chat pages. This file has blocked crawlers since at least September 2024, according to archived snapshots. However, robots.txt alone doesn't guarantee pages stay out of search results.

Both Google and Bing require an additional safeguard: a "noindex" tag embedded in individual pages or their response headers. When WIRED examined exposed Claude chat pages, none included this tag.

Google's documentation states the search engine will ignore robots.txt instructions if a page is linked from elsewhere on the internet and lacks the noindex tag. Bing's technical guidelines similarly recommend using both robots.txt and noindex tags together.

At the time of reporting, Bing still displayed approximately 612 results when searching for Claude share links, while Google had cleared its results. Many individual chats flagged in a Reddit post identifying the problem had been deleted.

Why it matters

This incident reveals a critical gap between AI companies' privacy promises and their technical implementation. Users sharing Claude conversations likely assumed Anthropic's systems would keep those links from becoming broadly discoverable through search engines. The exposure demonstrates that even sophisticated AI labs can overlook fundamental web security practices, potentially eroding user trust in chatbot privacy at a time when these tools handle increasingly sensitive queries.

AI labs blocking each other

The situation highlights broader tensions in the AI industry around web crawling. Anthropic, Meta, and OpenAI all block competitors' web crawlers from accessing their chatbot websites through robots.txt files—the same mechanism that failed to keep Claude chats private.

Google spokesperson Ned Adriance told WIRED that indexing responsibility lies with Anthropic. "Neither Google nor any other search engine controls what pages are made public on the web," Adriance said. "We give site owners clear controls to decide whether pages can be crawled or indexed, and we always respect those directives."

This isn't the first time Anthropic has faced this issue. Last September, the company encountered similar criticism and defended its use of robots.txt to Forbes, despite no guarantee the approach would prevent search indexing.

Claude users concerned about privacy can review their shared chats by navigating to Settings > Privacy > Shared chats. The exposed pages still lack noindex tags, meaning they could reappear in search results.

These details were first reported by WIRED, with additional reporting by Andrew Couts.