← Back to Blog
AI & Filmmaking · June 14, 2026

The Uncanny Valley Isn't Closing, It's Just Moving

By Pranav Arya · PAFP · #ai-video · #opinion · #technique · #generative-ai

Every model generation gets pitched as finally closing the uncanny valley, and every generation does genuinely improve overall convincingness while the specific thing that trips a viewer's subconscious discomfort simply relocates to a new, subtler detail rather than vanishing. Worth tracking where that discomfort has actually moved to, because it changes what to watch for in current output.

Where the tell used to live

A couple of years back, the obvious giveaways were structural, warped hands, physically impossible motion, faces that shifted identity slightly between frames. Those specific failures are largely solved in current top-tier models, and it genuinely felt for a while like the uncanny valley was closing for good as those glaring errors disappeared one by one.

Where it's actually living now

The current tell, in my experience reviewing a lot of generated footage, is subtler: micro-expressions and eye movement that are technically correct but slightly too smooth, missing the small asymmetries and hesitations real human faces have constantly and unconsciously. It doesn't read as "wrong" the way a warped hand does. It reads as a faint, hard-to-articulate wrongness that a viewer often can't name but still feels, which is arguably a harder problem to solve precisely because it's harder to even identify clearly.

What this means practically

When reviewing generated footage for anything requiring a genuinely convincing human presence, I've retrained my own eye to specifically watch faces closely for that overly smooth quality rather than scanning for the older, more obvious errors I used to check for automatically. I expect this pattern to keep repeating, each generation solving the previous tell and revealing a new, subtler one, rather than a single generation arriving that fully closes the gap once and for all.

Pranav Arya is a Berlin-based filmmaker producing AI video content for brands and social media, alongside real-world event, brand, and fashion shoots worldwide. He also teaches photography and videography to aspiring creators. Get in touch to work together.