AI Takes Director's Chair: How To Make A Blockbuster Short Film Using ChatGPT And Other Tools

Get ready to witness the ultimate mashup of science fiction and artificial intelligence as we explore how to use AI-powered tools to create a short film that’ll leave Hollywood scrambling for the next big thing. 

What Happened: Tech YouTuber Matt Wolfe has created a buzz with his latest video, which showcases the power of AI-powered tools like OpenAI’s chatGPT and Midjourney. 

In the video, Wolfe demonstrates how he used a combination of eight different AI platforms to create a short sci-fi film from start to finish, except for the editing process. 

See Also: Can ChatGPT Whip Up A Tasty Dish From These Random Ingredients? Sorted Food Finds Out

Here’s a step-by-step breakdown of how Wolfe created a unique, creative, and full-of-possibilities movie using just AI: 

Step I: Create A Script — ChatGPT

Wolfe wanted to create a documentary-style short film without dialogues. He inserted a prompt asking chatGPT-4 to write a script for a 60-90 seconds movie about “an AI robot that saves civilization.”

Step II: Detailed Shot List — ChatGPT

While in response to the initial response, GPT-4 gave instructive shot descriptions, Wolfe asked for more detailed information about the objects or people included in each shot, which helped him during the next steps. 

Step III: Create Videos — Runway Gen 2

Runway Gen 2 is currently available in a closed beta version, but Wolfe used it extensively to create videos by inserting shot descriptions generated by chatGPT. Though it worked well, but Wolfe had to try several times for some descriptions to get the right outcome. 

Step IV: Still Images – Midjourney 

For some reason, Runway Gen 2 didn’t produce a result for a humanoid robot on a kind of “operating table,” so Wolfe used Midjourney to generate an image for the same prompt. Midjourney was also used to generate several other still pictures for the short film. 

Step V: Time Lapse — Genmo 

For one segment, Wolfe needed a shot of a time-lapse video – from a destroyed dystopian world to a more colorful and dynamic city life. For this purpose, he used Genmo and got the perfect result. 

Step VI: Adding Razzle Dazzle 

To make the still images look more realistic or have a slightly heightened dramatic effect, Wolfe used LeiaPix and Kaiber – the results will blow your mind. 

Step VII: Voiceover — Eleven Labs 

Now for the voiceover, the script for the narrator was generated by chatGPT in Step I, so for the voiceover part, Wolfe used Eleven Labs. He chose the accent, age and gender and adjusted strength and stability. The issue Wolfe faced here was the pace of the voiceover — something he worked on during the editing process. 

Step VIII: Background Music — Mubert 

The last AI platform Wolfe used for Mubert to generate a piece of dark and dramatic background music. Wolfe removed a watermark in the download copy of this music during the editing step. 

For editing and bringing the entire concept to a final place, the YouTuber used Davinci Resolve by Blackmagic Design. 

Watch the aforementioned steps in action and the AI-generated short film here: 

At the time of writing, the video had been viewed by over 52,000 people. 

View more Tech Hacks on Benzinga here. 

Read Next: YouTube Channel Freakin’ Reviews Gives ChatGPT A Whirl — And The Verdict Is Mixed