Amid a roaring AI wave, the presence of long-video platforms is both eye-catching and slightly subtle. Rather than saying this is an active strategic choice on their part, it seems more like a form of passive defense driven by the push of technology.

Nowadays, the main battleground for long videos has been steadily losing ground, facing the risk of being overrun. As the most critical surge period of the year for long drama breakout hits, this year’s summer season hasn’t brought much good news. According to Jiemian Entertainment statistics, as of now, only Youku’s (Jiǔ Mén) and (Qiáo Chǔ) have exceeded 10,000 in heat, clearing the breakout-hit kicking line. The basic foundation of long dramas is visibly loosening.

AI has had the greatest and most direct impact on the film and television industry. AI’s technical support has made short dramas more powerful, competing for users’ attention. According to QuestMobile data, the stickiness of short-drama users has already surpassed online video, with average monthly usage time per user nearly doubling in reverse. On the other hand, AI keeps breaking through the barriers of the film and television industry, and narratives like “a vocational school graduate used 10 days and 3,000 yuan to create a viral AI short film” keep being written and spread repeatedly.

AI short dramas and AI comic dramas are gradually becoming mainstream directions for film and television adaptation, but even if platforms increase their investment, it is still hard to shake the user mindset and usage inertia that Red Fruit Free Comic Drama has established over the long term. According to QuestMobile data, among the top 10 short-drama apps by monthly active users, Red Fruit Free Short Drama still has nearly a 30-fold gap compared with other comic-drama apps.

However, increasingly mature AI tools have also freed content direction from being confined to short content under 10 minutes. AI films have moved from a conceptual product to a tangible visual entity. Not long ago, the first AI feature film to receive a (online drama distribution license), (Qitan: Zhi Ren Du Huangxu), went live on iQIYI; within less than a month, the AIGC feature-film series (Fushengmeng) from the “Soul Ferry” IP also began rolling out online.

At first glance, long-video platforms seem to have found their breakthrough direction, but under rapidly changing technical conditions, whether this competitive advantage can secure a stable future is still debatable. In exchanges with multiple AI super-creators and other industry insiders, Jiemian Entertainment found that insufficient talent supply and changes in channel logic are the two major hurdles long-video platforms must face. These hurdles are forcing long-video platforms to quickly recalibrate their position under the rules of the new era and reshape the content ecosystem.

Before the AI era, the path to becoming a film and television professional was almost fixed: through standard academic training or long-term on-set experience, and then being differentiated into different roles according to the needs of industrialized production. But the technological equalization brought by AI has made the standards for film and television talent much more generalized.

“Mx-Shell” had no formal background in film, television, or gaming, and through self-study of music, photography, and AI video tools, gradually created (Zombie Cleaner), which attracted Hollywood; “Feifei Fei,” who had been a model for 7 years, is now learning AI tools and created (The Girl Who Was Laid Off), making “Fang Taozi” the hottest AI performer.

The rise of this batch of “wild creators” has nearly rendered the traditional talent-evaluation system speechless. An AI project负责人 once told Jiemian Entertainment that when hiring, they value familiarity with AI tools more, as well as whether candidates have AI works; having experience in the film and television industry or formal training in a film and television major is actually secondary, and may even make them more cautious, because these people’s thinking patterns may still be stuck in the past and find it hard to quickly adapt to the new creative flow.

This change is especially tricky for long-video platforms. The director, screenwriter, and producer resources accumulated in the past no longer hold an absolute advantage in the AI era. To adapt to the current pace of content production, long-video platforms must expand their talent pools and reach those “new creators” who operate outside the mainstream spotlight and master the work language of the new generation.

The industry defines these “new creators” as “AI super-creators,” meaning super creators with the ability to produce AI works. Faced with the most in-demand talent in the AI era, platforms have all pulled out the stops, launching AI creator recruitment programs, hackathons, and other activities, using prize pools of millions, free tokens, and traffic support to attract creators.

For example, iQIYI launched the AI Film and TV Creation Camp, and together with Baodexi initiated the “Baodexi · iQIYI Film Theater” AIGC commercial film and TV venture program; the “HorsePower · AI Image Competition” co-hosted by Youku’s parent company and HappyHorse, along with the launch of the “HappyHorse AI Film and TV Practical Camp”; Tencent Video’s global creator-oriented full-chain incubation plan for overseas premium AI film and TV projects.

But in this talent competition, long-video platforms are not the only participants; large-model companies, AI tool companies, MCN companies, and other platform-side players are all involved. For example, Bilibili’s “AI Creation Contest,” Yuewen Group’s “Spark Plan,” and LibTV’s “Skill Super-Creator Incentive Program.” The melee among multiple forces has made the talent competition increasingly intense.

However, an industry insider told Jiemian Entertainment that the earliest entities to gather AI super-creators were mostly large-model companies, and the top AI super-creators in the market had basically already been “harvested” by them, with exclusive cooperation relationships established long ago.

More importantly, among all platform choices, advanced large models and easy-to-use AI tools are what really attract AI super-creators. AI super-creator Johnson told Jiemian Entertainment that what attracts AI super-creators first and foremost is definitely strong tool capability, “because when a general large model is updated, AI super-creators have priority access to internal testing.” Although long-video platforms also have self-developed AI tools, limited model capabilities make it hard for them to create a siphoning effect based on hardware strength alone.

If they can’t snatch AI super-creators but can still firmly control distribution channels, they can at least stay proactive upstream in the industrial chain. But now the traffic environment and distribution logic downstream are no longer the same.

The biggest disruption brought by the rise of short dramas to the film and television industry is that it has opened up a brand-new path for users to access content. Platforms like “Douyin, Kuaishou, RedNote, and Bilibili” have taken on this content demand, and short-form content has become users’ first choice today. According to QuestMobile data, users’ consumption of video content is evolving into a direct competition between “short videos + short dramas” and “long videos.”

In the early stage of AI development, limited by large-model capabilities, AI works could usually only be tens of seconds or a few minutes long. Such short content was naturally suited to spread on “Douyin, Kuaishou, RedNote, and Bilibili” social platforms, which also meant that users’ first impression of AI works, their interaction habits, and their mental anchors were all firmly built on short videos and social distribution, having long formed a kind of behavioral inertia.

By contrast, the distribution logic of long-video platforms is purpose-driven consumption: users enter with the clear expectation of chasing a series or watching a variety show, and they actively consume based on subjective judgments about IP and celebrities. The algorithmic recommendation logic of “Douyin, Kuaishou, RedNote, and Bilibili” is more aligned with fragmented content consumption habits, seamlessly embedding AI works into the feed based on users’ interest tags and achieving efficient matching where “content finds people.”

More importantly, the advantage of short-video platforms is not just that the subsidies they offer are larger, but that an AI super-creator’s account can simultaneously handle advertising, livestreaming, e-commerce, fan payments, platform revenue sharing, and IP collaborations. A single video may not make money, but as long as the account keeps accumulating followers, the creator has a chance to achieve a self-sustaining loop. This ecosystem of “the account is the brand, and the content is the shelf” is something long-video platforms find hard to satisfy.

For this reason, the split of traffic revenue and business opportunities from works is an important factor for AI super-creators when considering platform binding. As a result, iQIYI also announced its latest revenue-sharing rules at a recent creators’ conference, noting that the short-drama track has completely broken through format barriers: live-action short dramas, AI short dramas, and comic dramas all follow a unified revenue rule; creators of exclusive works receive a 100% full revenue share; non-exclusive works have a stable 70% revenue-share ratio; and creators who join iQIYI’s AIGC Creative Center are given commercial orders every month.

However, some AI super-creators told Jiemian Entertainment that market demand for AI commercial orders is now large, and commercial orders are not an especially attractive condition in themselves, since every platform can provide decent commercial-order resources. In other words, the advantage long-video platforms have on the monetization side for creators does not form a moat.

But that doesn’t mean long-video platforms have lost their appeal to AI super-creators across the entire upstream and downstream chain. In interviews, these AI super-creators unanimously mentioned the resource advantage of long-video platforms: besides commercial orders, long-video platforms can provide rich opportunities for film and television creation, including AI creations for any related business, such as but not limited to MVs and post-production for film and TV.

From another perspective, in this AI race, the long-video platforms’ opponents have never been social media platforms or MCN companies; rather, it is about rebuilding the content ecosystem for the next stage in this process and figuring out their own rules.

“AI is just a tool; content is what determines life or death.” Johnson told Jiemian Entertainment that the requirements for AI super-creators have changed compared with before: the focus is no longer on technical proficiency, but on content innovation.

The AI works that currently attract the public usually rely on excellent content rather than eye-catching technology. Not long ago, a hit AI short film on Bilibili, (My Greatest Sin in This Life Was Writing People as Demons…), gained tens of millions of views, and the most-liked comments in the section were more about the original story (Soushen Ji) and the historical event referenced, the Yellow Turban Uprising.

Audiences have begun to scrutinize AI works the same way they scrutinize traditional film and television productions, which is a sign that the content trend is shifting toward premiumization. And “premiumization” is precisely the content direction most familiar to long-video platforms that have gone through the IP wars.

Although AI has disrupted the content ecosystem, it has not fundamentally shaken the logic of content production as a whole, and the standards for high-quality content have not changed. The creative experience and aesthetic judgment accumulated by long-video platforms through years of cultivating “premium content” have instead become scarce resources at this point.

Although technological equalization has erased certain boundaries of talent standards, it cannot fill the natural gap in aesthetic perception, innovative ability, and narrative style. AI super-creator Dudu told Jiemian Entertainment, “The more powerful the model becomes, the more it ultimately returns to tradition. Why do we insist on working with traditional director teams? Because years of accumulated film and television techniques, creative inspiration, story structure and rhythm, aesthetics, lighting and cinematography... these things really cannot be simply replaced by technology. If you don’t have it in your head, you just can’t make good work.”

And those AI short dramas and AI short films can also be a convenient way for long-video platforms to break away from the previous, cumbersome process of judging creators’ ability through review. An AI short-drama director once told Jiemian Entertainment that after her AI short drama was released, a long-video platform contacted her to discuss cooperation on a live-action mid-length drama project.

The past success of long-video platforms came from organizing scarce professional content and then delivering it to audiences through centralized channels. AI film and television is doing the opposite: it first decentralizes production capabilities so that a large number of new creators can enter, and then tools, algorithms, and communities filter out a small number of works from massive experimentation. Neither system is absolutely superior; they simply require completely different entry points and rhythms.

Therefore, whether long-video platforms can keep up with this round of the AI race does not hinge on which day they produce an AI film worthy of promotion, nor on where a single model evaluation ranks. The real exam is whether platforms can find their place in the AI era and maintain control over content. Whoever can do that may be able to go from being merely a distributor of AI works to once again becoming the maker of industry rules.