I
Agentic Intelligence · Infomly

Anthropic just exposed 612+ Claude conversations to Google. Your employees' shared chats are indexed.

AI-Assisted Content — Produced with AI assistance and human editorial review. Learn more
612 Claude conversations showed up in Google and Bing search results this weekend.

Not because of a hack.
Not because of a breach.

Because Anthropic forgot the "noindex" tag.

Your employees who used Claude's share feature — to send a draft to a colleague, to save a research thread, to collaborate on a document — created public URLs that search engines indexed and made discoverable by anyone.

What WIRED found in those chats: medical billing dashboards, cryptocurrency wallet keys, children's contact information, internal company documents, legal discussions about ethical violations.

This is not a privacy bug. This is a product design failure with enterprise consequences.

The pattern: Anthropic tells users "anyone with the link" can view shared chats. Users interpret that as "only people I send this to." But the URLs are plain public HTML — no authentication, no access control, no noindex tag. The moment one lands in a Reddit post, a Discord message, or a public GitHub gist, it's indexed.

This happened before. In September 2025, Forbes reported the exact same issue. Anthropic said they used robots.txt to block crawlers. Google's own documentation says they ignore robots.txt if a page is linked from elsewhere and lacks the noindex tag.

Nothing changed.

If your organization allows employees to use Claude, audit your exposure today.

Go to Settings > Privacy > Shared chats > Manage. Review every shared conversation. Delete anything containing company data, credentials, or client information.

Then update your AI usage policy: shared links on consumer AI platforms are not secure collaboration channels. Period.

SOURCE: https://www.wired.com/story/private-claude-chats-exposed-in-google-and-bing-search-results
VERIFIED: WIRED, TechCrunch, 404 Media, Malwarebytes
SIGNAL: This is the third time Anthropic has had this issue since September 2025. Enterprise CISOs treating AI share features as internal collaboration tools just learned their data was publicly indexed for months.
💬 Consultation · Got questions? Talk to an expert →
Enterprise AI Impact — filtered for signal, not noise The AI briefing CTOs read before their morning meeting 3 minutes. Zero fluff. Only what moves the needle. $5/mo — your cheapest competitive edge
Subscribe — $5/mo

0 Comments

No comments yet. Be the first.