Головні звуки українського ІТ. Вгадаєш всі? 👉

Dialogues with Claude hit Google: how to protect your chats

Because of Claude's sharing feature, search engines started indexing chats that appeared to be private, leading Google to remove these pages from its search results.

Leave a comment
Dialogues with Claude hit Google: how to protect your chats

Because of Claude's sharing feature, search engines started indexing chats that appeared to be private, leading Google to remove these pages from its search results.

Users began to notice that dialogues with Claude appeared in search engines such as Google, especially if they added site:claude.ai/share to the search query, writes Neowin.

The main reason these conversations are publicly available at all is because of the built-in sharing feature: it’s designed to create a static web snapshot of the entire chat session. When someone clicks the “Share” button with the “Create Public Link” option enabled, Claude generates a public URL with a unique set of characters. Posting this link on social media or in an open developer forum allows anyone to view the conversation.

As Reddit users noticed, these pages did not have the proper HTML noindex tags or HTTP headers that would tell search bots not to index the content. Anthropic relied mostly on a simple robots.txt file to block the directory of common links from crawlers. However, search bots typically ignore such files if external sites link directly to these URLs.

Among the leaked files were legal strategy notes from practicing attorneys, seed phrases for recovering private cryptocurrency wallets, backend source code with encrypted database access data in plain text, and highly personal discussions of medical symptoms. Confidential corporate financial models and internal product development plans that employees copied directly into chat windows were also made public.

When you share Claude conversations, public links created by Free, Pro, and Max users are fully public. However, for Team and Enterprise users, these links are protected by organizational authorization, so only logged-in colleagues can view their content.

If you suspect you may have shared a conversation with Claude that wasn't meant to be public, you can unshare it by going to your account settings menu, opening the privacy tab, going to your list of shared chats, and clicking the Unshare button.

OpenAI found itself in a similar situation before — in August 2025, when Google began indexing thousands of conversations from ChatGPT.

At the time of writing, Neowin confirms that Google has already begun removing links to Claude's chats from search results, although some of them can still be found on Bing.

Anthropic has introduced the Claude Opus 5. The company says it is close to the Fable 5 but half the price. What is known about the new model
Anthropic has introduced the Claude Opus 5. The company says it is close to the Fable 5, but half the price. What is known about the new model
On the topic
Anthropic has introduced the Claude Opus 5. The company says it is close to the Fable 5, but half the price. What is known about the new model
Anthropic has updated the Claude voice mode: now users can choose between the Opus Sonnet and Haiku models
Anthropic has updated the Claude voice mode: now users can choose between Opus, Sonnet and Haiku models
On the topic
Anthropic has updated the Claude voice mode: now users can choose between Opus, Sonnet and Haiku models
Claude Code gets a built-in browser. Now AI can read websites, click buttons, and fill out forms
Claude Code gets a built-in browser. Now AI can read websites, click buttons, and fill out forms
On the topic
Claude Code gets a built-in browser. Now AI can read websites, click buttons, and fill out forms
Read the country's main IT news in our Telegram
Read the country's main IT news in our Telegram
On the topic
Read the country's main IT news in our Telegram
Also Read
Roosh запускає нову освітню платформу AI HOUSE CLUB для ML/AI-спеціалістів та дата сайнтистів. Розповідаємо, як подати заявку та чому навчатимуть
Roosh запускає нову освітню платформу AI HOUSE CLUB для ML/AI-спеціалістів та дата сайнтистів. Розповідаємо, як подати заявку та чому навчатимуть
Roosh запускає нову освітню платформу AI HOUSE CLUB для ML/AI-спеціалістів та дата сайнтистів. Розповідаємо, як подати заявку та чому навчатимуть
Головоломка киянина Quadline перемогла на фестивалі інді-ігор Google Play
Головоломка киянина Quadline перемогла на фестивалі інді-ігор Google Play
Головоломка киянина Quadline перемогла на фестивалі інді-ігор Google Play
Як нейромережі бачать вільну та незалежну Україну? Тест dev.ua
Як нейромережі бачать вільну та незалежну Україну? Тест dev.ua
Як нейромережі бачать вільну та незалежну Україну? Тест dev.ua
Нейронні мережі для генерації зображень бачать світ по-своєму, їхню логіку зрозуміти часом зовсім неможливо. Але таки хочеться. На честь Дня Незалежності України редакція dev.ua вирішила провести невеликий експеримент. Ми задали чотирьом різним нейронним мережам п’ять однакових запитів: «прапор України», «День Незалежності України», «український Крим», «перемога України» та «українці». Отриманими результатами ми ділимося з вами нижче.
Харківська художниця намалювала новий дудл для Google на День Незалежності України
Харківська художниця намалювала новий дудл для Google на День Незалежності України
Харківська художниця намалювала новий дудл для Google на День Незалежності України

Have important news to share? Message our Telegram bot

Key events and useful links in our Telegram channel

Discussion
No comments yet.