The AI Dubbing Atlas
- 01
- 02
- 03
- 04
- 05
- 06
- 01
- 02
- 03
- 04
- 05
- 06
The Pipeline
The field was moving faster than any single map could keep up with. This is the first survey.
AI dubbing is not a future prediction. Around 95% of all video content in the world is never dubbed into any new language due to cost — and that constraint is dissolving in real time. The Atlas is an attempt to give the industry a shared vocabulary before the vocabulary invents itself without us.
Each map in this series surveys a different stratum: the technical pipeline, the economic logic, the trust infrastructure, and the roles that don't exist yet. The goal is not a forecast but a reference — the kind of document you annotate, argue with, and return to.
This first survey covers the pipeline cross-section: six layers from source audio to final deliverable, and an account of where machines now handle the routine and where humans still make the difference that matters.
Machines can speak. But speaking is not the same as moving someone.
How this publication works
Every statistic is sourced and cited. Where a figure appears in the text, the methodology is described. Contested numbers are flagged.
The Atlas surveys the AI dubbing pipeline as it exists today — not as it is marketed. Claims about capability are tested against published outputs, not vendor demos.
No generated text is published as editorial. Where AI tools assist in research or structuring, that is disclosed. The voice in this publication is a human voice.
This field moves quickly. Each map carries a publication date and a revision log. When the territory changes, the map changes.
The 2050 industry doesn't exist yet. We're building it — and it will be built by people who cared enough to draw the first maps.
The AI Dubbing Atlas
Draft section / dubformer.ai/atlas