China's Platforms Are Betting on AI Interactive Content

Something odd is happening to short video in China. The category that ate the internet is running out of room to grow. As of June 2026, the active penetration rate of short video in the country hit 92.6 percent, while the entire mobile internet base, 1.282 billion monthly active users, grew just 1.2 percent year over year. When almost everyone is already watching, the fight stops being about more users and turns into a fight over what they watch next.

Four of the biggest platforms have answered that question with the same word: interaction. Over the past few months, Bilibili, Xiaohongshu, Kuaishou, and Douyin each moved into what the industry calls AI interactive content, and the shape of the bet tells you a lot about how Chinese platforms think about the next format.

From a personality test to a platform feature

The spark was small. In April 2026, a Bilibili creator called Q-Rou made an AI-generated personality test that happened to land on exactly the right internet joke at the right moment. It spread across the whole Chinese web, and the related videos on Bilibili have passed 5.7 million views.

Bilibili noticed fast. Two months later it started testing Toy, a publishing surface that turns a local interactive web page into something you can open and play directly inside the app. Within about a month, users had racked up more than 8.5 million interactions, and creators had uploaded over 300 works: games, interactive stories, tools, tests, reports, and picture books.

Xiaohongshu followed in July, rolling RED Skill out to everyone and testing a Vibe Coding tool. RED Skill holds reusable AI skill packs that a user copies and runs elsewhere, while Vibe Coding produces small web apps that can be opened right inside a note. Kuaishou announced an AI interactive content co-creation program. Douyin opened an interactive space plaza where creators showcase their own works, from branching stories to casual games.

Why platforms want this so badly

The obvious reason is engagement. A short video is passive. You watch it, it ends, you swipe. An interactive piece asks you to do something, which keeps you on the surface longer and gives the platform more signals about what you like. That is valuable to an advertising business that has to keep showing growth even when the audience has stopped growing.

There is a second reason that is less discussed. Interactive content is a way to get creators to produce things a recommendation feed cannot. A feed rewards whatever performs, which pushes everyone toward the same handful of formats. Interactive pieces are harder to clone, so the platform ends up with a supply of distinct work that a competing feed would struggle to match. The creator, in turn, gets something they own rather than another swing at an algorithmic lottery.

The catch nobody has solved

The enthusiasm runs ahead of the economics. A viral interactive piece has a short life. People play it once, share it, and move on, which is a poor fit for the kind of repeat engagement platforms and brands actually want. Turning a burst of attention into a sustainable revenue stream has proved hard, and few, if any, interactive works have closed the loop between popularity and money.

There is also a quality problem hiding under the novelty. Plenty of the early works are thin: a personality quiz, a jokey calculator, a two-choice story. These spread because they are quick and shareable, not because they are good. If the format settles into gimmick territory, audiences will tire of it the way they tired of earlier one-note trends, and platforms will be left explaining a feature that never found a business.

What creators actually get

Step back from the platform strategy and the appeal to a creator is easier to see. Making a video takes time: scripting, shooting, editing, and a dozen rounds of revision. Making an interactive piece can take less, because the interaction carries part of the experience and the audience supplies the rest. A creator with an idea and a laptop can ship something playable in a weekend instead of assembling a crew.

That lower barrier is the same force reshaping short drama and image generation across the Chinese internet. Cheap generation tools mean the cost of a first attempt drops toward zero, and when attempts are cheap, people attempt more. Some of those attempts become the next viral format. Most do not, but the platform does not care which ones succeed, as long as something on its surface does. It is a lottery with a very low ticket price, and creators keep buying in.

The risk is that the same cheapness floods the surface with interchangeable works. The personality tests and two-choice stories that spread early are easy to clone, and once everyone makes one, none of them stand out. The interactive format may survive its novelty phase only if creators start building things that are genuinely hard to copy.

There is also a language problem that keeps coming up in these discussions. Interactive works travel across platforms differently than video does, because a game or a test can be embedded almost anywhere while a video is bound to its feed. That portability is a strength for the creator and a concern for the platform, which would rather keep the interaction inside its own app. Bilibili, Xiaohongshu, and Douyin are all building their own hosting surfaces in part to keep that traffic from leaking out. The fight over interactive content is quietly also a fight over who owns the link.

Why this matters outside China

Chinese platforms tend to hit these walls first, because their market matured first. That makes them a preview. If short video saturates everywhere else the way it has in China, the same question arrives: when passive consumption stops growing, what replaces the swipe?

Interactive AI content is one candidate, and it comes with a built-in dependency. Almost every one of these works is made with generative tools, which means the format's ceiling is tied to how good and how cheap those tools are. Cheap generation is what makes a thousand experimental interactive pieces possible. Better generation is what makes them worth playing twice.

The platforms are not betting that interactive content will replace short video. They are betting that it will sit alongside it, catching the audience that has grown bored of watching and wants to poke at something. Whether that audience is large enough to build a business on is still an open question. But the fact that four major platforms are asking it at the same time says the era of pure watching is closer to its end than its beginning.