ChatGPT and the future of YouTube

ChatGPT is a large language model with an interactive chat interface. What this means is, you can ask the AI to write for you, with some details about what you’d like written, what you want to include, and what form it should take. This has implications for school essays, work productivity (getting unstuck from writer’s block), but also, the future of YouTube videos.

Just as there’s a huge debate over AI generative art, the meaning of art, the value of art, and the copyright implications of generative art having been fed with unlicensed art used without the permission of the creator / copyright owner, there’s a huge debate about what AI generative text means.

For example, it’s possible that the last school essays written by humans have already been written. Today, Chat GPT can return a plagiarism checker-resistant essay that’s mediocre, but passable. As the models improve, it’s possible that better essays result.

People suggest that it’s a great tool for getting unstuck from writer’s block. (ChatGPT was not used in the creation of this post.) But the difference between using it as a tool to springboard off of, and using the writing it creates directly, while noticeable today, may become slim in the future.

There are people who are using it for entire emails. Any time in the past when you had to worry about being understood could be gone. Differences between languages, between non-native speakers, and just plain difficulty with mastery of grammar fade away when the AI re-writes your email so you can be better understood.

YouTube

“Hey guys, remember to like, subscribe, and hit that bell for notifications!”

Making Youtube videos has traditionally been hard. You have to script, you have to film, create header images, animate, and so on. It’s a lot of work. And on top of that, it takes ages to build an audience.

TikTok user @cadenboof2 has figured out a hack to make YouTube videos quickly. He has posted just 2 TikTok videos and 28 shorts on YouTube, but his shorts on YouTube have millions of views. None of them appear to be automated with his process, although he has other posts named, “Doing homework with Artificial Intelligence” and “I sold feet pics with Artificial Intelligence”. The posts are thoroughly in the “side hustle” genre.

The process

  1. write the script at chatGPT using a prompt like, “script a 1500 word script about making money”
  2. Copy and paste the script into pictory.com using Script to Video
  3. Then go to Synthesia.io and choose a character they offer. Paste in the whole script.
  4. then use a free video editor like wisecut.video to put the synthesia.io video on top of the pictory.com video
  5. get someone on upwork to make a thumbnail
  6. use upwork to get someone to do the whole process

Why this fails

The problem is, it’s a race to the bottom on quality. Is the crappy automated movie good enough to watch? Is it cheap / easy enough to make?

Great, there will be a flood of cheap, crap videos. People will try and avoid these, so engagement will get crushed. People have short attention spans, and need something interesting or controversial in the first part of the video to keep watching.

If the content is uninteresting and is easily identified as automated by the viewer, no one will watch, there will be no money in it, and no one will make them. That’s the situation today. But generative AI has been rapidly evolving.

Today’s failure is tomorrow’s possibility

What’s boring and insufficient today could be much improved in the next year, or the year after that. People are already writing books with ChatGPT. It’s writing emails for people, correcting for tone, spelling, and non-native speaker concerns. What’s to say we won’t have generative AI video that works soon enough?

Leave a Reply

Your email address will not be published. Required fields are marked *