"Is your video account still in place?
Many people are stuck at the first step in making YouTube videos based on AI. It's not that the tools are difficult to use, it's that I don't understand how to do it. There are now many cases of using AI to make Short Video on TikTok. Some people rely on AI to edit and produce 50 videos a week, and their accounts become popular and take off directly. The key is to use AI on ideas and don't mess around.
Let's start with how to choose tools. There are two types of AI video tools on the market today. One type is fully automatic, such as directly inputting text to generate video. This type is suitable for beginners, but the disadvantage is that it is too template. The other type is semi-automatic, such as using Coze to build a workflow. This type takes some time to configure, but can make differentiated content. Recently, I have seen many bloggers using the combination of Runway ML+Coze, and the results are good.
How to operate it specifically? Take it in three steps. The first step is to find a benchmark account and disassemble it. Don't underestimate this step. Many people start using tools directly, but as a result, no one looks at what they make. Spending two hours studying peer video scripts, rhythms, and transitions is much better than blindly trying. For example, the recently popular "Travel Vlog" account has a fixed formula for each video: scenery shots must appear in the first 3 seconds, with 3 story points interspersed in the middle, and suspense at the end.
The second step is to use AI to generate material. Nowadays, AI video tools all support keyword generation. But using the default template directly, the video looks like it was written by AI. You can try the "reverse operation method": first manually shoot a set of basic materials, and then use AI to expand the capacity. For example, I took a meal scene and asked AI to generate meal videos from different angles and different scenes. This not only ensures authenticity, but also allows mass production.
The third step is to optimize the details. Videos generated by AI tools usually have two problems: inaccurate stuck points and stiff dubbing. The solution is simple. Stuck points are manually fine-tuned using clipping and reflection, and dubbing is re-recorded using Descript's AI repair function. The free version of these two tools can be used, which is much better than directly generated by AI.
Many people ask whether AI can completely replace manual labor? Of course not. But it can greatly reduce the production threshold. For example, when making a Short Video in the past, it might take a long day to find the material and edit it. Now with AI, these processes can be compressed to half an hour. Using the time saved on content planning is the value AI should play.
It was interesting to see a case recently. There is a blogger who uses AI to generate 5 different styles of videos every day, and then looks at the data to pick up hits. Ordinary people can learn this kind of test-based style of play. The key is to find your place. For example, if you cook delicious food, use AI to generate the production process of various dishes. If you are a knowledge person, you will generate explanation videos from different angles.
Nowadays, YouTube's recommendation algorithms are increasingly focusing on completion rates. AI video has a natural advantage: it can quickly test different starts. For example, for the same theme, use AI to generate 10 versions with different starting points to see which data is the best, and then use this model firmly later. This data-driven approach is much more reliable than creating based on feeling.
Tell me a point that many people ignore. No matter how powerful AI tools are, they cannot solve the problem of topic selection. Nowadays, content homogeneity is serious, and using AI to make videos requires more attention to differentiation. For example, in the same AI painting tutorial, some talk about tool usage, some talk about creative skills, and some talk about commercial monetization. Which path your account is going to take, think clearly before you start.
By the way, don't use AI as a lazy artifact. Real experts use AI as "assisted creation". For example, use AI to generate a first draft of a script and then rewrite it using your own experience. This not only improves efficiency, but also maintains personal style. Many popular videos on the market today are created by this "human-machine collaboration" model.
Competition in the video industry will only become more intense. If you rely on pure manual work slowly, you will be eliminated sooner or later. However, relying entirely on AI for mass production is also easy to fall into the trap of low-quality content. Finding a balance point is the long-term way. For example, I use AI to make 3 test videos and 1 piece of high-quality content that I cut myself every day. In this way, we can not only measure data, but also accumulate masterpieces. "
