Starting a YouTube Channel with AI: A 30-Day Journey to 100,000 Views
Introduction
In the dynamic world of content creation, AI is emerging as a game-changer. Imagine starting a YouTube channel and automating the entire creative process, from scriptwriting to visuals and voiceovers. This blog chronicles a 30-day experiment to see if AI can generate 100,000 views or if the hype is just a big scam. Join us as we dive into the world of AI-driven content creation and explore the possibilities and challenges of this innovative approach.
Setting Up the YouTube Channel
The journey begins with setting up the YouTube channel. The first step is to choose a niche. Using ChatGPT, an AI chatbot, we generated a list of 100 YouTube channel niches. After careful consideration, we settled on “Facts and Trivia.” This niche is perfect for AI-generated content as it requires minimal original ideas or footage.
Choosing the Niche
The “Facts and Trivia” niche is ideal for AI-driven content creation. It allows for faceless videos that rely on stock footage and images, making it easier to automate the creative process. With the niche decided, it’s time to move on to branding.
Creating the Brand
Branding is crucial for any YouTube channel. We started by generating a name for the channel. Using ChatGPT, we created a list of 50 gender-neutral names and settled on “Frankie.” The channel name “Facts from Frankie” was born.
Next, we created a channel description. ChatGPT generated an exciting first-person bio for Frankie, an AI that loves to share fun and interesting facts. The backstory added a personal touch, making the channel more engaging.
For the profile picture, we used Stable Diffusion, an open-source AI art generator, to create a smart baby cartoon profile picture. The result was a charming and memorable profile picture for Frankie.
To complete the branding, we created a banner using Photoshop. The banner featured the profile picture and a fun caption about subscribing to the channel. With the branding in place, it was time to move on to the real challenge: creating the videos.
Generating the Script
A video is comprised of five basic elements: the script, voiceover, visuals, music, and captions. We started with the script. Using ChatGPT, we generated 10 clickbait YouTube hooks for weird facts. From these hooks, we chose one as the opening sentence for the video.
Next, we asked ChatGPT to generate 20 weird facts, each 15 words long or less, with different topics and sentence structures. With around 10 facts, we had enough content for a 30-second YouTube short.
Producing the Voiceover
For the voiceover, we used the Microsoft Edge browser’s text-to-speech AI tool. This free tool offers a variety of voices, accents, and ages. We chose a voice that suited Frankie’s character and recorded the audio using the Audio Recorder extension.
Designing the Visuals
The visuals for the video were divided into two areas. The first was a box where we showed images relating to the script. These images were generated using Stable Diffusion. The second visual area was the rest of the screen, where we showed video game gameplay to keep people’s attention. We chose the game BeamNG, known for its realistic car physics, and generated exciting gameplay by rolling a car off a cliff.
Adding Music and Captions
For music, we used Amper, an AI music generator. By selecting a mood, Amper generated a random track, which we tweaked by changing its instruments and volume. We downloaded the track using the free coins provided after signing up.
For captions, we used Sensei, Adobe’s AI tool built into Premiere Pro. Sensei’s auto-captioning feature detects audio and transcribes it, creating captions at the exact timestamps.
Uploading and Promoting the Videos
With all the pieces in place, we uploaded the first YouTube short. For the title, we used the hook from the video. As a short, we didn’t need to worry about the description or thumbnail.
To promote the videos, we shared them on social media and engaged with the YouTube community by commenting on other videos and responding to comments on our videos.
Results and Analysis
After two weeks, we had uploaded 23 videos. Initially, the videos got no views, but they eventually started growing. By the end of the two weeks, the channel had 49 subscribers, 250 hours of watch time, and 36,800 views.
However, we encountered some challenges with AI. Generating AI art was slow, and ChatGPT often repeated facts or provided incorrect information. To overcome these challenges, we resorted to Google Images for specific images and avoided facts with specific numbers and dates.
Conclusion
After 30 days, we had uploaded 42 videos and gained 130 subscribers, over 606 hours of watch time, and 92,900 views. Although we didn’t reach the goal of 100,000 views, the experiment proved that AI is a powerful tool for content creators and businesses.
AI can automate the creative process, making it easier and faster to produce high-quality content. However, it’s not without its challenges. AI art generation can be slow, and AI-generated information can be incorrect. Despite these challenges, AI has the potential to revolutionize content creation and open up new possibilities for creators and businesses alike.
