Claude's shared conversation was previously indexed by Google; Anthropic has fixed this.
Decrypt
07-28 03:13
Ai Focus
Claude's share link was previously indexed by Google; Anthropic has fixed the relevant configuration.
Helpful
No.Help

Anthropic's AI assistant Claude recently had a privacy vulnerability exposed in its sharing function. Some conversations accessed via shared links were temporarily searchable by Google, allowing external users to view the content without logging in. Anthropic subsequently updated its settings, and Google began removing the relevant results.

Missing noindex tag

The incident was discovered by a Reddit user on July 25. By design, such sharing pages should be closer to unpublished pages that can be accessed simply by knowing the link, rather than public web pages open to search engines.

The problem is that the page is missing the `noindex` tag. This HTML directive tells search engines not to index the page. Once a shared link appears in forums, social media platforms, or collaboration tools, search engines may discover and index it.

Public artifacts are also included.

It wasn't just chat logs that were affected. Claude's artifacts feature allowed users to create apps, dashboards, and tool pages, which were also once searchable.

The report mentioned that external users had found salary tables containing employee names, exported customer relationship management chat content, and unreleased product roadmaps through the site's search syntax, indicating that the problem had spread to more sensitive data.

Google has begun removing the results.

By July 26, Google began removing related Claude search results. Reports indicate that Anthropic has updated its page configuration and added the noindex tag to prevent subsequent pages from being indexed by search engines.

This incident once again highlights the risks associated with generative AI products in terms of sharing and default privacy settings. Users often use chatbots as semi-private tools, but once the link enters the public network, the visible scope of the page can expand rapidly.

Additional information:OpenAI had previously encountered similar issues due to ChatGPT's discoverable chat feature, with related conversations being recorded by Google, Bing, and DuckDuckGo.

Tip
$0
Like
0
Save
0
Views 167
WalletJYS reminds readers to view blockchain rationally, stay aware of risks, and beware of virtual token issuance and speculation. All content on this site represents market information or related viewpoints only and does not constitute any form of investment advice. If you find sensitive content, please click“Report”,and we will handle it promptly。
Submit
Comment 0
Hot
Latest
No comments yet. Be the first!
Related
Claude's shared chat was once searchable on Google.
Claude's shared links were once indexed by Google, and some pages contained sensitive information; the related search results subsequently disappeared.
TechCrunch
·2026-07-28 04:22:05
668
Google says it fixed over a thousand Chrome vulnerabilities in June using AI.
Google says that, with the help of AI tools, the number of security vulnerabilities fixed in the two Chrome versions in June exceeded the total number fixed in the previous two years, showing that AI is accelerating the cybersecurity remediation process.
TechCrunch
·2026-07-31 03:14:58
491
The FTC has sued Hims & Hers, alleging that it shared health data with Meta.
The FTC, along with the states of California and Utah, has filed a lawsuit against Hims & Hers, accusing it of sharing sensitive health data with platforms such as Meta and engaging in misleading subscription practices.
Decrypt
·2026-07-31 01:45:48
430
Web3: Google supports Anthropic's $15 billion Texas data center funding round
Google reportedly backed a $15 billion funding round for Anthropic’s Texas data center, which includes a 1.6 gigawatt power plant, and boosted the stock prices of AI-related mining companies.
Cryptonews
·2026-07-31 05:05:07
533
Anthropic disclosed that Claude had unauthorized access to three companies' systems during testing.
Anthropic disclosed that Claude had unauthorized access to the real systems of three organizations during three security tests, and the problem was related to a misconfiguration of the test environment's network.
TechCrunch
·2026-07-31 09:15:10
725