AI dubbing and localization: how studios are translating games in 2026
Voice cloning, automatic lip-sync: AI voice engines are cutting the cost and turnaround of localization. Voice actors and purists aren't all on board.
This article contains affiliate links. If you buy through one of them, PicGamers may earn a commission. It never changes the price you pay and never influences our scores.
Localizing a AAA game into fifteen languages used to cost several hundred thousand dollars and months of studio time. Generative AI voice engines are changing that math, at the cost of a real fight within the dubbing industry.
What the tech can actually do today
The pitch: a voice actor records a performance in one language, then a voice-cloning model generates the other versions from that vocal fingerprint, with automatic lip-sync remapping the facial animation to the new language instead of the other way around. Some publishers already use it for secondary languages (Polish, Dutch, Brazilian Portuguese) that used to simply ship subtitled for lack of budget. The output still reads as synthetic on emotionally heavy scenes, but largely passes unnoticed on functional dialogue and background NPCs.
Pushback from voice actors
The real friction point isn’t quality, it’s the contract: several actor unions are now demanding explicit clauses on how a cloned voice can be reused after the initial recording, with additional pay for every new generated language. Some studios sidestep the issue by limiting AI use to languages they’d never have dubbed with real actors anyway — an argument that softens the blow, but hasn’t convinced everyone.
Our take
The tech doesn’t yet replace careful dubbing on lead roles, but it has already filled a real gap: languages that simply weren’t dubbed before. Worth watching on the contract side as much as the technical one.
