I have a theory.
The loneliest place in the world isn’t an empty room.
It’s a crowded train where nobody looks up.
Those sentences grabbed me. It was the spark for my latest AI filmmaking experiment — a short poetic film about loneliness, connection, and the forgotten magic of two strangers becoming a story.
The strange thing about our modern world is that we have never had more ways to connect.
We carry tiny computers in our pockets that can reach almost anyone on Earth.
And yet sometimes we are surrounded by people and feel completely invisible.
That contradiction fascinated me.
So I decided to make a video about it.
Not a lecture.
Not a warning about technology.
A reminder.
Maybe we don’t need more followers.
Maybe we just need someone to look up and say:
“I’m really glad we met.”
2/9
The idea: making loneliness visible
The first challenge was translating an emotional idea into images.
AI is very good at making things that look impressive.
It is not always good at making things that feel human.
There is a difference.
A beautiful image can still be emotionally empty.
For this project, I needed a visual symbol.
I found it in the character of a little girl.
Not because children are magical fantasy creatures. (Although, having raised kids, I can confirm they are occasionally magical chaos goblins.)
The reason I chose a child was simpler:
Children still look.
They notice.
They say hello.
They are curious before they learn all the reasons not to be.
The little girl became the emotional center of the story.
The adults represent the world we sometimes become.
The child represents the world we started with.
Lab note:
The biggest surprise in AI filmmaking is that the “character” is often not the character.
It is the idea behind the character.
The AI is not just creating a person.
It is helping create a symbol.
3/9
Step 1: Building the visual identity
Continuity is one of the hardest problems in AI video.
Anyone who has tried creating multiple AI shots knows the pain.
You generate a beautiful character.
Then the next image arrives.
Different hair.
Different face.
Different clothing.
Congratulations. Your character has apparently traveled through three alternate universes between scenes.
For this project, I created a strong reference image:
A cheerful 4-year-old Caucasian girl with vibrant natural red hair, sparkling blue eyes, rosy freckles, soft porcelain skin, and a genuine joyful smile.
Her clothing was kept consistent:
A cream-colored linen dress with soft blue embroidered flowers and a light beige cardigan.
The goal was not just visual continuity.
The goal was emotional continuity.
Every time she appeared, she needed to feel like the same little person.
Not a collection of AI-generated pictures.
A character.
Lab note:
AI still struggles with memory.
That means the filmmaker has to become the memory.
The human has to hold the story together.
4/9
Step 2: Creating the world around her
The poem begins in isolation.
So the first half of the video needed visual distance.
Crowded places.
People everywhere.
Nobody connecting.
I created shots of:
• A crowded commuter train where everyone is looking down at phones
• A cafe full of conversations where someone sits alone
• A college campus where thousands of people pass each other
• A city street filled with people but lacking eye contact
The interesting challenge was making loneliness without making everything depressing.
I didn’t want a dark dystopian future.
I wanted Tuesday afternoon.
Because that is where the real loneliness lives.
Not in science fiction.
In ordinary moments.
The person sitting three feet away.
The person waiting in line.
The stranger who might have become your friend if somebody had simply said hello.
5/9
Step 3: The music decision
The music choice was critical.
I needed something recognizable.
Something people already carried emotional memories for.
The answer was obvious:
“What a Wonderful World.”
The song has been part of our cultural memory for decades.
But instead of using the original recording, I used a piano and cello cover.
Then I converted the key from F major to D major to better match the little girl’s singing voice.
The result was exactly what I wanted.
A familiar melody.
A new emotional context.
The first verse is sung by the AI-generated child character, creating a moment where innocence and nostalgia meet.
Lab note:
One of the coolest things about AI creativity is that you can create combinations that would normally require a production studio.
A child performer.
A cinematic portrait.
A custom musical arrangement.
A complete short film.
All created from my little corner of the world.
6/9
Step 4: Bringing the voice to life
The narration uses an AI recreation inspired by the legendary voice quality of Sir Laurence Olivier.
The reason was not imitation for imitation’s sake.
The poem needed a storyteller.
Someone who sounded reflective.
Someone who felt like they had lived long enough to understand that moments matter.
The voice needed warmth.
Not a commercial announcer.
Not a documentary narrator.
A person sitting across from you saying:
“Here is something I’ve been thinking about.”
That distinction matters.
AI voices are improving quickly, but direction still matters.
The tool can speak.
The filmmaker has to decide what it means.
7/9
The creative stack
This project was built with a combination of AI and traditional editing tools:
• Midjourney — character design, cinematic stills, environment creation
• ElevenLabs — AI narration and vocal performance
• WAN AI – AI lip sync workflow — bringing the little girl’s singing performance to life
• FLOW MUSIC – Music processing tools
• AUDACITY – adjusting the instrumental track to fit the vocal key
• Kdenlive — editing, timing, transitions, and final assembly
The final video runs approximately 2 minutes and 57 seconds.
The narration is about 1 minute and 36 seconds.
The singing sequence adds another 25 seconds.
The remaining time allows the instrumental music and visuals to breathe.
Because sometimes the most important part of a film is the moment after the words stop.
8/9
The real lesson
This project reminded me of something important about AI filmmaking.
The technology is not the story.
The technology is the paintbrush.
The camera.
The editing room.
The orchestra.
The human still has to decide what deserves attention.
The most important prompt was not the Midjourney prompt.
It was the question:
“What do I want people to feel?”
The answer was simple.
Look up.
Notice people.
Leave room for beautiful accidents.
The next person who changes your life might not arrive with a dramatic entrance.
They might just be the stranger walking past you today.
Lab note:
The best AI results often come from happy accidents.
A slightly different expression.
A surprising composition.
A moment where the machine misunderstands your idea and accidentally reveals a better one.
That is where the collaboration becomes interesting.
The AI creates possibilities.
The human chooses meaning.
9/9
The final thought…
I started this project thinking I was making a video about loneliness.
Somewhere along the way, it became a video about hope.
That seems to happen often with creativity.
You begin with a question.
You end with a discovery.
Maybe the world is not as disconnected as it appears.
Maybe there are millions of people quietly waiting for someone else to make the first move.
Maybe the next great conversation is one hello away.
And maybe AI, for all its strange imperfections, can help us remember something deeply human:
We were never meant to simply be seen.
We were meant to be known.
I am still learning this new craft.
Still experimenting.
Still breaking things.
Still occasionally yelling at AI like it personally betrayed me.
But somewhere between the prompts, the pixels, the music, and the mistakes…
there is a little bit of magic.
And that is why I keep creating.
Steve Teare
video alchemist
TerminallyBored.Monster
Palouse, Washington USA