We’ve crossed the uncanny valley, now we need to teach the machines to act

Originally published by The Drum, January 8, 2026

I recently did something that would have been science fiction five years ago and prohibitively expensive five months ago. I cloned myself.

Using the latest generation of AI avatar technology, I created a digital twin. I fed the machine footage of my face, my voice, and my mannerisms. When the render came back, the result was technically flawless. It looked like me. It sounded like me. It even had that specific way I tilt my head when I’m making a point.

The technology has officially crossed the uncanny valley. The visual glitches are gone.

But as I sat there watching Digital Colin, I realized we’re facing a new problem. The avatar looked human, but it was acting like a newsreader. Technically perfect. Completely boring.

This is the trap everyone’s falling into right now. We’re so impressed that the technology works that we’ve forgotten our actual job: to entertain the audience.

Welcome to the boredom valley

LinkedIn and corporate inboxes are being flooded with AI avatar content. Almost all of it follows the same format: a perfectly lit, unmoving face staring directly into the lens, delivering a monologue for two minutes straight.

It’s efficient. It’s scalable. And it’s unwatchable.

Real cinema doesn’t work like this. We rarely hold on a single shot of a talking head for more than a few seconds. We cut away. We use reaction shots. We change the angle. Film theory 101.

The current wave of avatar content feels robotic, not because the render is bad, but because the filmmaking is lazy. We’ve confused a tech demo with a film.

Directing your digital self

When I set out to use my own avatar for a recent project, I made a conscious decision to reject the monologue format. If I wanted the audience to connect with the content, I had to treat the avatar not as a piece of software, but as a difficult actor who needed directing.

But first, I needed decent source material. I shot proper stills of myself – professionally lit, interesting wardrobe, plain background. Not some AI-generated office nonsense that screams: ‘I didn’t leave my bedroom.’ If you’re going to clone yourself, at least make it look intentional.

We started with the story, not the tech. We built something that required visual dynamism. We didn’t just let the avatar talk – we made it perform within a wider cinematic context. The avatar became a narrator, not an anchor.

Then we did what you’d do with any footage that wasn’t quite landing: we cut away. High-end b-roll illustrated the points being made. Sound design underscored the emotional beats the avatar’s face might miss. We played with pacing, allowing for pauses that felt natural rather than algorithmic.

This is where craft comes back into play. To make an AI feel human, you have to use the language of cinema to compensate for its limitations. You have to understand that Digital Colin can’t improvise or react to the room, so the edit has to do that work for him.

The result? People stopped talking about the tech and started talking about the story.

Taste is the new barrier to entry

This experience made something very clear. The technical barrier to entry has evaporated. Anyone with a credit card can generate a photorealistic human. The new barrier to entry is taste.

If you treat an AI avatar like a text-to-speech engine, you’ll get a robotic performance. But if you treat it like talent, everything changes. You have to write differently for an avatar. You have to write for the edit.

We’re moving from a world of ‘capturing’ performance to ‘designing’ performance. The best creatives of the next decade won’t just be the ones who can write the best prompts. They’ll be the ones who understand human psychology well enough to inject imperfections back into the machine. They’ll know that a stutter, a pause, or a jarring camera angle is what makes something feel real.

The idea matters more than ever

There’s a fear that avatars will replace human connection. I see it differently. I see them as the ultimate removal of friction.

By digitizing my likeness, I removed the logistical nightmare of production. I didn’t need lights, a crew, or a studio booking. I could iterate on the script in real time. I could fix a stumble in the delivery without a reshoot.

This didn’t make the creative process colder. It made it more honest. Without the distraction of the production circus, there was nowhere to hide. If the script was bad, the video was bad. If the idea was dull, the avatar couldn’t save it.

We’re entering an era where the only thing that limits us is our imagination, not our schedule. The machines are ready. Now we just have to give them something interesting to say.