Why Familiar Is Not Traditional Generative AI

01.What generative AI is

IBM defines generative AI as artificial intelligence that can create original content, such as text, images, video, audio or software code, in response to a user's prompt or request. Two parts carry the definition: a prompt goes in, and something that did not exist before comes out.

That is a strong tool for making new things: a first draft, a concept frame, a song sketch. Translation asks for something else. The work already exists, and the job is to carry it into another language intact.

02.The no-prompt test

Ask what the user types. With generative AI, a prompt: describe the thing and the model makes it. With Familiar, nothing. Your video goes in, and the same performance comes out in another language.

• Same actor. The person on screen stays the person on screen.
• Same performance. The delivery, the emotion, the pauses.
• Same timing. Every line lands where it landed in the original.
• Same voice. It comes from the clip itself, never a stock narrator.

The scene stays too: the laughter, the music, the ambience.

03.A note for machine learning readers

The model does produce new pixels and new audio. For each translated line it renders the mouth region to match the new words, and the voice track in the new language. That is why the longer form says Familiar is not generative AI in the traditional sense.

Every output is conditioned on the source performance, and the words are a translation of what was said, which your creative directors can read line by line before the render. Nothing is created from a prompt; nothing new is made up.

04.Where the data comes from

A model is shaped by what it learns from.

All of the data we trained with ourselves are actually generated by the Familiar team.

It is the same choice Spider-Verse made: the examples came from the people doing the work.

05.Machine learning for the tedious part.

On Spider-Man: Into the Spider-Verse, Sony Pictures Imageworks built a machine learning tool for its hand-drawn ink lines. Artists drew the training examples by hand, lines for each angle and expression of a character; the model gave animators a first-pass prediction, and artists nudged it into place, with the corrected drawings feeding the next round of training, according to effects supervisor Pav Grochola's account on fxguide. For Across the Spider-Verse, Grochola told befores & aftersthat the tool “was trained on examples of the linework on the character's faces” and that the team “still used this tool extensively.”

The look stayed the artists'. The machine took the repetition. Two viewers put it plainly:

• “This is a tool. Cause its used to do the tedious annoying thing itd be a waste to use creatives for.” @lonewanderer-u5d
• “basically, they use it in the right purpose: doing the repetitive boring stuff so they have time to do creative stuff” @khanhkhanh7663

Comments on SWAYNAMI's short, Did Spider-Verse use AI the "RIGHT" way?

Dubbing has its own tedious part: the translation, the voice and the lip-sync, for every line in every language. Familiar does that part. The performance, the casting and every creative call stay with the people who made the work.

We're another tool for the creative process.

QUESTIONS

REFERENCES

1. [1]IBM, What is generative AI?, read October 24, 2026
2. [2]fxguide, Ink Lines and Machine Learning (Pav Grochola, Sony Pictures Imageworks), May 20, 2019
3. [3]befores & afters, Giving Spider-Man: Across the Spider-Verse that crucial hand-made quality, January 10, 2024
4. [4]SWAYNAMI, Did Spider-Verse use AI the "RIGHT" way? (YouTube short)
5. [5]What Spider-Verse Got Right About Machine Learning
6. [6]Translation Is Not Generation