Login is required to view the full text, and overseas manufacturers have also started to pick up bad habits.

In recent days, a hot topic has emerged in the global overseas internet circle: Google AI Mode is siphoning off web traffic. The gist is that users are increasingly getting used to obtaining answers directly from the summaries and conversations generated by Google AI, instead of clicking links to visit specific websites.

Websites that have had their traffic cut off naturally cannot tolerate this. Even Reddit, which once reached a data cooperation agreement with Google in the past, has "turned against" it. Reddit has decided to adjust its operation strategy, no longer expecting traffic from Google search to return to previous levels. It plans to gradually shift to an access mode dominated by logged-in users, requiring visitors to create an account and log in before they can browse relevant content.

In the future, users who are not logged in may only be able to see the titles of Reddit posts, and they will need to log in if they want to reply, like, vote and so on. According to Reddit, the login mechanism can help Reddit build a more direct relationship with users and reduce its dependence on external search engines.

Logged-in users can provide more data on their interests and behaviors, making it easier for Reddit to carry out personalized recommendations and ad delivery. By restricting anonymous access and automated crawling, Reddit can better control the way content is used, and convert data authorization into commercial revenue.

Obviously, the strategy once adopted by Zhihu of "no full content access without logging in" has been directly copied by Reddit.

In fact, it is not that Reddit is deliberately making things difficult for non-logged-in users, the siphoning power of AI is simply too strong. Reddit just released its financial report for the third quarter of 2026, which shows that the number of daily active unique visitors in the United States has declined, dropping from 53.2 million in the previous quarter to 50.7 million in the third quarter. If search traffic continues to shrink, Reddit estimates that the number of global daily active unique visitors in the fourth quarter of this year may see a decline of less than double-digit percentage.

For information that originally required users to visit Reddit to find, Google AI can directly provide a summary at the top of the search page, so why would users bother to visit Reddit's website. Once websites no longer provide full content to visitors, crawlers of search engines cannot continue to obtain full content unless they ignore the robots protocol.

When Reddit's future is overshadowed, investors immediately voted with their feet, and its stock price plummeted by about 30% on October 30, marking one of the worst single-day drops since its listing. "All the gifts fate has bestowed come with a hidden price tag". The $60 million annual data licensing agreement Reddit reached with Google two years ago eventually made them pay the price.

The data provided by Reddit allows Google to continuously iterate its AI models, and the AI trained by Google makes users no longer need to visit Reddit. In a sense, Reddit can be said to have personally nurtured its own "grave digger". Now Reddit is preparing to stop its wrongdoings before it is too late, and stop the bleeding wound that has already started before the situation becomes irreparable.

Then the question arises: since both platforms force users to log in to browse content, is Reddit not afraid of repeating Zhihu's mistakes?

As early as May 2024, Zhihu suddenly implemented a rule on its web end that non-logged-in users could not view full content. But only 7 months later, in early 2025, Zhihu quietly lifted this restriction, allowing users to browse full content directly without logging in.

Zhihu lifted the restriction of "no full content access without logging in" because this measure does not fit the domestic internet ecosystem in China. In fact, the web ecosystem of China's internet has been deliberately undermined by major manufacturers, and the experience on the web end has been declining ever since the birth of the mobile internet. When the core way users access internet services is through mobile apps instead of web pages on browsers, Zhihu's practice of restricting visitors' access to web version services is doomed to be futile.

In contrast, thanks to earlier development and higher PC penetration rate, the overseas web ecosystem is not inferior to the app ecosystem, or even on a par with it. Today, the overseas web ecosystem is still full of vitality, which makes Reddit's measure of banning visitors from browsing full content effective. For this company that relies on user scale to attract advertisers and maintains platform activity through community content, retaining users is the only way to secure a future.

In fact, AI manufacturers have also realized that they cannot continue to squeeze content platforms. At the same time, OpenAI, Anthropic, Google and Microsoft jointly announced that they will adopt a new standard, with which content publishers can choose to block AI crawlers from accessing content on their own websites, and the standard can be integrated into the robots.txt file of the website.

If no measures are taken to stop the bleeding for internet content platforms, once other manufacturers start to follow Reddit's example, everything may become irreparable.

This article is from WeChat Official Account "San Yi Life" (ID: IT-3eLife), written by San Yijun, and published by 36Kr with authorization.