Last month, the first AI-made domestic theatrical film to receive a public screening license hit theaters.

But the response was underwhelming. Not only was the projected box office under one million yuan, but on social media, some viewers complained that the film's quality was "even worse than Niu Lai."

So I've deliberately omitted the film's title—no need to add fuel to the fire.

After all, as an early foray into AI's entry into the film industry, it is more experimental in nature. Given the current limitations of AI generation technology, no one can achieve perfection, and it is pointless to amplify and mock those shortcomings.

What we really ought to discuss is the somewhat bizarre situation that has emerged in the AI film field: moviegoers aren't excited about AI, AI commercialization doesn't need films, yet industry practitioners are all eagerly anticipating AI films.

Currently, the most important commercialization market for AI is not theatrical films at all. Data disclosed by Chinese Online shows that 55% of domestic AI token consumption goes toward AI short dramas.

Yet the market is still anticipating "the first AI film." After this AI film was released, the stock prices of the relevant production companies rose accordingly, and it was hailed as a landmark achievement in AI application implementation.

But the vast majority of audiences simply don't care. The few who went did so with the same mentality as checking out Niu Lai—purely out of curiosity, willing to sample even something terrible just to see for themselves.

The various reactions to AI films are quite telling. Some are frenzied, some indifferent, some opportunistic, some fleeing—this is precisely a unique contemporary spectacle created by the combined effects of the current transformation period across all industries and the AI transition period.

The facts are obvious: AI films are not yet mature, and the works currently shown in theaters fall far short of the quality standards of traditional theatrical feature films.

But what view to take—criticism or encouragement—the value judgment depends on how you understand it. Objectively speaking, AI films are in a stage of advancing from short dramas to feature films. The tech stack and industrial processes are not yet fully stable, and a production system capable of supporting a high-quality theatrical feature film has yet to be established.

If we use the mature genre film standards of the century-old film industry as the yardstick, then AI films don't even measure up to Niu Lai in terms of laughs and social value. But if we use the early stage of a new creative model as the criterion, then we can't be too harsh—after all, innovation requires allowing trial and error, and it would be even better if audiences weren't asked to pay for it.

That said, why, knowing it is immature, would AI films still be brought to theaters for public screening? Is the film industry trying to cash in again? It seems the film industry winter isn't cold enough—some people will surely suspect this. So it is necessary to briefly analyze the actual situation of AI films today.

To summarize in one sentence: the film industry's understanding of and research into AI actually far exceeds outside imagination—even beyond what some tech observers imagine. I have seen directors and screenwriters speak about AIGC, computing power, and GPUs with great familiarity, being quite knowledgeable about technology. Some AI super-creators have even transitioned directly into AI directors without any traditional filmmaking experience. However, the century-old film technology and industrial system will not automatically change course within two or three years of AI's impact. In reality, AI's entry into film keeps hitting walls.

For example, in pre-production creation, technical constraints are significant—filmmakers currently still need to accommodate technology.

In traditional film creation, the director first proposes creative intent—this is a director-centered or screenwriter-centered system—and then different professional teams find implementation solutions. But after AI joins filmmaking, the boundaries of model capabilities must be considered, which in turn limits the conception and creativity of the script.

The first AI theatrical film, because AI at the time could not stably portray delicate emotional changes between characters, steered its narrative toward civilization fables, sci-fi crises, and other themes more suited to technical conditions in order to avoid technical shortcomings.

Moving into the production and shooting process, film crews now add the position of AI director, responsible for translating the creative intentions of the chief director, assistant director, art director, and other team members into language that AI can understand.

However, in reality, the work content of an AI director cannot yet be clearly defined. How much to pay for this role and how to price it—the industry has no specific standards or figures. In short, this industrialized production process involving AI directors is also in a constantly changing and unstable state.

Next comes the distribution and screening stage. Currently, there is much uncertainty about whether AI films can be shown on the big screen. Industry insiders reveal that purely AI-generated narrative feature films may not be able to obtain the Dragon Logo—the public screening license issued by the National Film Administration—in the future. Policy uncertainty remains significant.

Even with the Dragon Logo, it does not mean the film can be stably delivered to theaters. Take the first AI theatrical film as an example: after the materials entered theaters, they encountered technical problems on the big screen such as insufficient resolution, dropped frames, and black screens. Even at the last moment before outputting the screening copy, individual shots were still being modified to meet screening standards.

In short, today's film + AI is still in a very awkward state. For instance, technology suppressing creativity is clearly putting the cart before the horse, because technology should serve creativity. Every link in production, distribution, and screening must respond to the impact of AI, but the accumulation from these responses seems useless because things may change again very soon.

Some practitioners say that the moment an AI film is completed, it is already a thing of the past—the AI version locked in by the film may already be behind the model capabilities of a few months later.

In such an uncertain cycle, if it were you, would you take a long-term approach to making AI films? And for roughly the next few years, AI films will remain in this "uncertain state." I suggest everyone watch with caution.

Unlike filmmakers desperately trying to embrace AI, AI companies and tech giants have an ambiguous, distant-yet-close attitude toward films.

On the one hand, AI companies use cinematic-quality visuals as a benchmark for their video generation models' capabilities. This leads model companies to often invite artists and directors to become their industry partners, promoting film as a key capability of their models. Tech companies like Netflix, Google, and ByteDance have also become headline sponsors of domestic and international film festivals, connecting film with their brands.

However, the capabilities tech companies draw from the film industry are rarely fed back into film itself. Instead, they are output to other content categories such as short dramas, medium-length dramas, and comic dramas.

This may also explain why the industrialization of AI films progresses slowly while AI short dramas and comic dramas explode with overwhelming momentum. Because tech companies invest vastly different amounts of effort and resources in the actual production of different content categories.

In scenarios like short dramas and comic dramas, AIGC video has denser supply and consumption, making it easier to form a commercial closed loop and allowing model companies to continuously make money.

AI commercialization does not rely on cinemas, so tech companies' embrace of film capabilities and abandonment of the film market is not hard to predict. After all, even the commercialization of film itself can hardly rely on cinemas anymore.

Audiences aren't even watching films anymore—so who is still waiting for AI films?

Films are not the main battleground for AI commercialization, so why is the capital market's reaction so strong? As soon as the first AI film launched, it repeatedly pushed related stock prices to the daily limit. Niu Lai was driven by retail investors to a million in box office, and now they have taken a liking to AI films—could retail investors be the true film fans?

In fact, just as the popularity of Niu Lai was never a victory for the film industry but merely an occupation of cinema space by internet meme culture, the investment and financing performance of AI films does not mean this category has consumption value—it is more about speculative value.

The reason is that film is a public content consumer product with national-level influence. For people hungry for high-quality AI, the previous national-level AI was embodied intelligence, which is now also yesterday's news. And the truly national-level AI apps, such as Doubao, Qianwen, and Work Buddy, are part of tech giants' businesses, with prices that already include traditional internet operations—not pure enough. Those pure AI applications and models are either unavailable for purchase or have already been speculated to astronomical price-to-earnings ratios of hundreds of times. So it is not that pure-blood AI is unaffordable, but that industry AI offers better value for money.

Among the many industry AI applications, film is undoubtedly the most direct AI implementation track, and also the one with the most influence and national recognition.

Although short videos, short dramas, games, and streaming have already diverted a huge amount of attention, and although cinemas are no longer the only place where everyone watches the same content together, the general public is still paying attention to films. Whenever a new film is released, media and self-media always launch countless discussions about its box office, story, cast, behind-the-scenes stories, and more. This makes AI films one of the AI application stories most easily understood and spread by the capital market.

Capital is speculating, while audiences are fleeing.

As mentioned earlier, AI technology has become the center of discussion in the film and television industry. More and more celebrities are selling their faces to AI companies, Hengdian has become a digital factory base for tech companies, and directors, screenwriters, and other creative teams are creating conditions to use AI even in processes that do not need it. Yet at the final consumer end, they say AI short dramas may not necessarily be inferior to films.

If filmmakers are all revolving around AI, focusing their attention on catching up with the technological wave, then why should audiences keep waiting for them and accompany them through trial and error?

This is a picture woven together by the technological transition period and the industry transformation period. Perhaps many people are waiting for AI films, but in reality, no one is waiting.

Practitioners are waiting for the impact and fluctuations of technology to pass and enter a new stable state; tech giants are waiting to absorb and digest the accumulation of the century-old film industry to nourish more profitable content categories like short dramas and comic dramas; speculators are waiting for life opportunities, finding one fleeting AI application after another in the ironclad cycle of technological revolution.

Perhaps one day in the future, the first mature AI film will appear—but no one will care anymore.