"There are millions of traffic hidden in this phone.
Short Video are now being made, and AI has helped a lot. It can not only write scripts, but also directly generate video clips. For example, if you ask AI to take a food tutorial, it will complete it for you in a few minutes.
But the problem is that many people can't use it well. Either the script is too fake, or the picture doesn't use your brain. In fact, AI is just a tool. The key depends on how you use it.
Let's talk about script first. Don't always ask AI to write the word "deep empowerment". Give an example directly, such as "Turn on a low heat when omelying eggs, and then pour the oil." AI understands, and the words written will be natural.
Talk about the picture again. AI can color colors and add special effects, but don't use them blindly. Many of the "retro filters" that TikTok has become popular recently are generated by AI with one click. But you have to choose the right scene and use retro shots to eat hot pot, and the audience will be confused.
There is also background music. AI can recommend music, but don't believe it all. Last week, a client approached me. His video AI was equipped with fast-paced music, but the picture was in slow motion. As a result, the number of views dropped by half.
So using AI to make videos, three points:
First, the script should be specific and not fake.
Second, the pictures must match and do not add special effects randomly.
Third, listen to music, not rely entirely on recommendations.
AI is getting smarter now. For example, the Coze platform can help you automatically generate video processes. You just need to make your needs clear, and it will set up a good framework.
But AI won't think for you. Last week, an Internet celebrity approached me. His video AI was done, but the comment area was full of questions,"Where is this?" In fact, he forgot to add the location description.
When making Short Video, AI can save time, but the details still depend on people. For example, AI can shoot the recently popular "contrast lens", but it's up to you to decide which lens comes first.
There is a recent trend: AI+ real people mixed cutting. For example, if you shoot yourself talking, AI generates an animation as the background. This kind of video has good traffic now.
But pay attention to the rhythm. When real people are talking, don't animate too fast, otherwise the audience won't be able to keep up. There was a case last week where a real person spoke for 5 seconds and the animation was cut three times, but no one finished watching it.
There are also subtitles. AI can be generated automatically, but don't trust it all. There was a video last week in which AI wrote "add some salt" as "add some money" and almost rolled over.
So use AI to make videos and don't be lazy. It is best to proofread subtitles yourself, especially sentences with numbers.
Nowadays, there are more and more AI tools. For example, the AI image cut can remove the background. But pay attention to the light. If the person is too dark, it will bloom when you pick it out.
There is also AI dubbing. Recently, many people have used AI to change their voices, such as changing male voices into female voices. But don't abuse it. Some platforms don't allow publishing.
Make Short Video, AI can help you, but don't rely on it all. For example, the recent popularity of "One Mirror to the End" cannot be done by AI, but it has to be photographed by real people.
AI can now automatically add subtitles. But don't use the default font, change it to bold for clarity. There was a video last week with the default italic subtitles, but the old man said he couldn't read clearly.
There are also transfers. AI can add cool transitions, but don't use it too fast. There was a case last week where the transfer was too fast and the audience thought it was stuck.
So use AI to make videos: let it do simple things, and you control complex things.
AI can now automatically color colors. But don't let it mess up. There was a video last week where AI turned the night view too bright and completely lost the atmosphere.
There is also AI to generate animations. For example, if you take a product introduction, AI can generate a 3D model. But pay attention to the proportion. Last time there was a customer, the cup generated by AI was bigger than a person.
Make Short Video, AI can help you speed up. For example, if you take 10 pieces of material, AI can automatically select the best one. But you still have to decide the key frame.
AI can now automatically edit. But don't let it cut. There was a video last week where the AI suddenly cut away before the laugh point, and the audience was confused.
So use AI to make videos: let it be resurrected, and you have to do the creative work.
AI can now automatically add tags. But don't believe everything. Last week, there was a video where AI added "delicious food", but the content was travel, but no one watched it.
Therefore, when using AI to make videos, it is best to choose your own label. Especially in unpopular fields, AI may not understand.
AI can now automatically generate thumbnails. But don't use too much. There was a video last week in which the platform directly restricted the streaming of large-character thumbnails generated by AI.
So use AI to make videos and adjust the thumbnails yourself. For example, adding a question mark is more attractive than a big word.
AI can now automatically analyze data. But don't believe everything. There was a video last week. AI said that it had a "high completion rate", but the comments were all complaints.
Therefore, when using AI to make videos, the data should be combined with the comments. Negative comments, in particular, may be the key.
AI can now automatically generate hot topics. But don't follow blindly. There was a topic last week that AI recommended, but the audience was not interested.
Therefore, when using AI to make videos, hot spots must be combined with account positioning. For example, if you are a food number, don't always talk about financial topics.
AI can now automatically optimize titles. But don't use too long. There was a title last week. AI generated a line of words, but no one clicked it.
So when making videos with AI, you have to change the title yourself. For example, adding a number is more attractive than a long sentence.
AI can now automatically generate scripts. But don't use too official. Last week, there was a script written by AI to "improve user experience" and the audience directly swiped it.
Therefore, when using AI to make videos, the scripts must be down-to-earth. For example, simply saying "this method saves money" is better than "optimise cost structure."
AI can now automatically generate questions and answers. But don't use something too fake. There was a question and answer last week. AI wrote "What do you think?", but the audience thought it was fake.
So using AI to make videos, questions and answers should be like chatting. For example, simply saying "Let's try it" is more natural than "Welcome interaction".
AI can now automatically generate summaries. But don't use something too big. There was a summary last week. AI wrote "Towards the Future", which stunned the audience.
So when making videos with AI, the summary should focus on the key points. For example, saying "see you tomorrow" directly is more direct than "look forward to meeting next time."
AI can now automatically generate suggestions. But don't believe everything. There was a suggestion last week. AI said to "shoot more beautiful women", but my account is in the technology category.
Therefore, when using AI to make videos, it is recommended to combine positioning. For example, if you are a Science and Technology account, don't always take pictures of beautiful women.
AI can now automatically generate ideas. But don't go too far. There was an idea last week. AI asked me to shoot "People Fly", but my equipment couldn't shoot it.
Therefore, when using AI to make videos, creativity must be combined with reality. For example, taking a small experiment on the table is more reliable than special effects.
AI can now automatically generate processes. But don't be too complicated. Last week, there was a process in which AI divided it into 12 steps, and the audience was stunned.
Therefore, the process of using AI to make videos needs to be simplified. For example, it is divided into 3 steps, which is clearer than 12 steps.
AI can now automatically generate templates. But don't be too rigid. There was a template last week where AI allowed fixed shots, but the audience was bored.
Therefore, when using AI to make videos, the template must be flexible. For example, inserting a close-up in the middle is better than fixing the whole process.
AI can now automatically generate feedback. But don't believe everything. There was a feedback last week. AI said that "viewers like slow motion", but my videos are fast-paced.
Therefore, when using AI to make videos, feedback must be combined with data. For example, if you have a high number of views, you really like it.
AI can now automatically generate categories. But don't be too broad. Last week, there was a category where AI was classified as "life", but the content was games.
Therefore, when using AI to make videos, the classification must be accurate. For example, simply saying "game skills" is more attractive than "life".
AI can now automatically generate keywords. But don't be too unpopular. Last week, there was a keyword. AI recommended "quantum mechanics", but the audience didn't understand it.
Therefore, when using AI to make videos, the keywords should be public. For example, using "money-saving techniques" has higher traffic than "quantum mechanics".
AI can now automatically generate recommendations. But don't be too complicated. There was a recommendation last week. AI asked me to shoot 10 topics, but the audience couldn't remember it.
Therefore, when using AI to make videos, recommendations should be focused. For example, choosing 1 topic is better than 10 topics.
AI can now automatically generate titles. But don't exaggerate. There was a headline last week that AI wrote "Shocking the World", but the content was average.
So when making videos with AI, the title must be true. For example, simply saying "this method is easy to use" is more reliable than "shocking the world."
AI can now automatically generate covers. But don't be too messy. There was a cover last week, and AI added 10 words, which the audience couldn't see clearly.
Therefore, when using AI to make videos, the cover should be concise. For example, putting a product picture is more attractive than words.
AI can now automatically generate scripts. But don't be too official. Last week, there was a script written by AI to "improve efficiency", which the audience thought was fake.
Therefore, when using AI to make videos, the scripts should be colloquial. For example, simply saying "this method is faster" is more natural than "improving efficiency."
AI can now automatically generate thumbnails. But don't be too vague. Last week, there was a thumbnail image with small AI-generated characters that the audience couldn't see clearly.
Therefore, when using AI to make videos, the thumbnail should be in large characters. For example, using "must-see" is more attractive than small words.
AI can now automatically generate transitions. But not too fast. There was a transition last week. The AI cut too fast and the audience thought it was stuck.
Therefore, when using AI to make videos, the transition should be moderate. For example, 0.5 seconds is smoother than 0.1 seconds.
AI can now generate subtitles automatically. But not too much. Last week there was a caption, AI in italics, old people can not see clearly.
So when making videos with AI, the subtitles should be in bold. For example, bold is clearer than italics.
AI can now automatically generate toning. But not too bright. There was a color adjustment last week. AI adjusted the night view too bright and lost the atmosphere.
Therefore, when using AI to make videos, the color must be natural. For example, a darker one will look better than a brighter one.
AI can now automatically generate dubbing. But don't be too fake. There was a dubbing last week. The AI's voice was too mechanical, and the audience acted.
So use AI to make videos, and dubbing requires real people. For example, recording yourself is better than AI.
AI can now automatically generate templates. But don't be too dead. There was a template last week where AI allowed fixed shots to make the audience bored.
Therefore, when using AI to make videos, the template must be flexible. For example, inserting a close-up in the middle is better than fixing the whole process.
AI can now automatically generate categories. But not too wide. Last week, there was a category where AI was classified as "life", but the content was games.
Therefore, when using AI to make videos, the classification must be accurate. For example, simply saying "game skills" is more attractive than "life".
AI can now automatically generate keywords. But not too cold. Last week, there was a keyword. AI recommended "quantum mechanics", but the audience didn't understand it.
Therefore, when using AI to make videos, the keywords should be public. For example, using "money-saving techniques" has higher traffic than "quantum mechanics".
AI can now automatically generate feedback. But don't believe everything. There was a feedback last week. AI said that "viewers like slow motion", but my videos are fast-paced.
Therefore, when using AI to make videos, feedback must be combined with data. For example, if you have a high number of views, you really like it.
AI can now automatically generate suggestions. But don't go too far. There was a suggestion last week. AI asked me to shoot "people fly", but my equipment couldn't shoot it.
Therefore, when using AI to make videos, it is recommended to combine it with reality. For example, taking a small experiment on the table is more reliable than special effects.
AI can now automatically generate ideas. But don't be too big. Last week, there was an idea that AI could "improve the user experience", which stunned the audience.
Therefore, when using AI to make videos, the creativity must be specific. For example, simply saying "this method saves money" is better than "improving the user experience."
AI can now automatically generate processes. But don't be too complicated. Last week, there was a process in which AI divided it into 12 steps, and the audience was stunned.
Therefore, the process of using AI to make videos needs to be simplified. For example, it is divided into 3 steps, which is clearer than 12 steps.
AI can now automatically generate questions and answers. But don't be too fake. There was a question and answer last week. AI wrote "What do you think?", but the audience thought it was fake.
So using AI to make videos, questions and answers should be like chatting. For example, simply saying "Let's try it" is more natural than "Welcome interaction".
AI can now automatically generate summaries. But don't be too big. There was a summary last week. AI wrote "Towards the Future", which stunned the audience.
So when making videos with AI, the summary should focus on the key points. For example, saying "see you tomorrow" directly is more direct than "look forward to meeting next time." "
