一条叫《早餐奶奶》的AI短视频，在抖音上拿了2.5亿播放、1509万点赞。
你猜它的创作者是什么背景？专业导演？MCN机构？大厂设计师？都不是。答案是：一个之前没做过短视频的普通人，用AI工具一个人做出来的。
这不是孤例。2026年上半年，抖音的AI内容生态正在经历一场爆发——涌现出9条破亿视频、20个半年涨粉超50万的账号。参与「OiiOii创作者计划」的0基础创作者们，只用了1个多月，累计播放量就达到了12亿。
注意，是累计12亿，不是单人。但即便如此，这个数字也足够惊人——它意味着AI内容创作的门槛，已经低到了一个前所未有的程度。
一人军团模式的规模化
传统短视频的生产模式是什么样的？一个团队——编剧、拍摄、剪辑、运营，少则三五人，多则十几人。产能有限，成本不低，能不能出爆款还要看运气。
AI时代的模式是：一个人，就是一支军团。
AI工具已经能覆盖短视频生产的几乎所有环节：文案生成、脚本拆分、画面生成、数字人口播、配音配乐、字幕、封面、标题、标签……以前需要好几个岗位配合的工作流，现在一个人加上AI工具就能搞定。
更可怕的是迭代速度。传统团队做一条视频，从选题到发布可能要三五天。AI创作者呢？一天产出十几条不是什么难事。数量级的差异，直接改变了内容生产的游戏规则——以前是精雕细琢碰运气，现在是快速试错、数据驱动、批量生产。
当然，量多不代表质高。9条破亿的背后，可能是几百条几千播放的炮灰。但这恰恰是AI内容生产的优势：试错成本几乎为零。一条没爆？没关系，下一条。十条没爆？第二十条可能就爆了。传统团队试错十次的成本，AI创作者可以试错一百次。
内容产业的供给侧革命
这件事最大的影响，不是多了几亿播放，也不是多了几个网红账号。而是内容产业的供给侧，被彻底改写了。
以前内容行业最大的瓶颈是什么？是人。好的创作者是稀缺资源，培养一个成熟的内容创作者需要好几年。MCN机构的核心竞争力，就是签下最多的好创作者。
AI把这个瓶颈打破了。0基础的人，经过简单培训就能上手做内容，而且产能还不低。这意味着内容供给会出现爆炸式增长——以后不缺内容了，缺的是注意力。
供给爆炸之后，会发生什么？
首先是平均质量先降后升。一开始大量低质量AI内容涌入，平台上充斥着同质化的东西。但很快，用户审美疲劳了，纯粹靠数量堆的打法会失效。能活下来的，还是那些有独特审美、有个人风格、有真实表达的创作者——AI只是他们的放大器。
其次是内容的个性化程度会大幅提高。当生产内容的成本趋近于零，给每个人定制专属内容就变得可行了。以后你刷到的每条视频，可能都是专门为你生成的。
最后是传统MCN模式会受到冲击。一个人就能干一个团队的活，为什么还要签MCN分成？当然MCN也有资源和经验的优势，但它们必须转型——从经纪人模式转向服务模式，给AI创作者提供工具、培训、商业化对接。
12亿播放只是一个开始。AI内容创作的浪潮才刚刚起来，接下来的一两年，我们会看到更多让人惊掉下巴的数据和案例。
但有一件事是不变的：工具再强，最后拼的还是创意和审美。AI可以降低门槛、提高效率，但它不能替你想出来要表达什么。真正好的内容，永远是人脑里出来的东西。
明天见。
An AI short video called "Breakfast Grandma" got 250 million views and 15.09 million likes on Douyin.
Guess what the creator's background is? A professional director? An MCN agency? A designer from a big tech company? None of the above. The answer: an ordinary person who'd never made short videos before, made it by themselves with AI tools.
This isn't an isolated case. In the first half of 2026, Douyin's AI content ecosystem is going through an explosion — 9 videos breaking 100 million views, 20 accounts gaining over 500,000 followers in half a year. Zero-experience creators participating in the "OiiOii Creator Program" hit a cumulative 1.2 billion views in just over a month.
Note: that's 1.2 billion cumulative, not per person. But even so, the number is staggering — it means the barrier to AI content creation has dropped to an unprecedented level.
The One-Person Army Model Goes Scale
What's the traditional short video production model? A team — screenwriter, filming, editing, operations — three to five people at minimum, sometimes over a dozen. Limited output, high costs, and whether you get a hit still depends on luck.
The AI era model is: one person, one army.
AI tools already cover almost every环节 of short video production: copy generation, script breakdown, image generation, digital human presenters, voiceover and music, subtitles, thumbnails, titles, tags… a workflow that used to require several different roles can now be handled by one person with AI tools.
What's even more striking is the iteration speed. A traditional team might take three to five days from concept to publish for one video. AI creators? Cranking out a dozen a day is nothing. An order-of-magnitude difference in quantity directly changes the rules of the content game — before, you refined and polished and hoped for luck; now, it's fast testing, data-driven, batch production.
Of course, more volume doesn't mean higher quality. Behind those 9 viral 100M-view videos, there might be hundreds of flops with a few thousand views each. But that's exactly the advantage of AI content production: the cost of failure is nearly zero. One didn't go viral? No problem — next one. Ten didn't work? The twentieth might.
The Supply-Side Revolution of the Content Industry
The biggest impact here isn't a few billion more views, or a handful of new influencer accounts. It's that the supply side of the content industry has been completely rewritten.
What used to be the biggest bottleneck in the content industry? People. Good creators are a scarce resource. Cultivating a mature content creator takes years. An MCN's core competitive advantage is signing the most good creators.
AI breaks that bottleneck. People with zero experience, after simple training, can get started making content — and their output isn't low. This means content supply will explode — soon there won't be a shortage of content; there'll be a shortage of attention.
After supply explodes, what happens?
First, average quality drops then rises. At first, a flood of low-quality AI content pours in, platforms are full of homogeneous stuff. But soon, users grow tired of it, and the pure-volume approach stops working. What survives are creators with unique aesthetics, personal style, authentic expression — AI is just their amplifier.
Second, content personalization will increase dramatically. When the cost of producing content approaches zero, customizing exclusive content for each person becomes feasible. In the future, every video you scroll past might be generated specifically for you.
Third, the traditional MCN model will be disrupted. If one person can do the work of a team, why sign with an MCN and give up a cut? MCNs of course have resources and experience advantages, but they'll have to transform — from talent agency model to service model, providing tools, training, and monetization support for AI creators.
1.2 billion views is just the beginning. The wave of AI content creation is just getting started. Over the next year or two, we'll see even more jaw-dropping data and cases.
But one thing stays the same: no matter how powerful the tools, the final competition is still creativity and taste. AI can lower barriers and boost efficiency, but it can't think of what to express for you. Truly good content always comes from the human mind.
See you tomorrow.
AI can lower barriers and boost efficiency, but it can't think of what to express for you.
— The Dawn Vision Editorial Desk
OiiOii · Douyin · AI creation · one-person army · content supply side · 1.2B views · zero-experience creators · AI new media
Sources · 信源 Sources
本文基于 Dawn Vision 认知引擎处理的 6 个源信号生成，经编辑部人工审核。素材来源：刺猬公社、新榜、爱范儿、NGA玩家社区。
This article was generated by the Dawn Vision cognitive engine processing 6 source signals, with human editorial review. Sources: Hedgehog Commune, NewRank, ifanr, NGA.
