Published rates put professional dubbing at roughly $20 to $50 a finished minute, before casting and studio time, while AI dubbing vendors sell the same job as a same-day subscription task and claim up to 92% less turnaround time. For a brand running one film across the GCC and South Asia, that difference decides how many markets the budget reaches.
We deliver films in English, Arabic, Urdu and Chinese, so the question arrives on most briefs. The useful answer is not which technology wins. It is which parts of a campaign each one belongs to.
What AI dubbing does well
- The long tail. Cutdowns, explainers, product walkthroughs and internal films, where the job is comprehension, not persuasion
- Volume across markets. One master into several language versions in a day instead of a week per language
- Voice continuity. Cloning keeps a familiar voice across versions, which matters when a brand has used the same read for years
- Testing. Trying a script in three dialects before committing a booth day to the winner
Where it breaks
Close-ups. The moment a face fills the frame, lip sync carries the performance, and dubbing of any kind has to be cut around, not pasted over. Humour and idiom break next, because a joke that lands in English rarely survives a literal translation into Arabic. And emphasis goes last: a synthetic read is even, and advertising depends on the uneven line, the pause, the one word that is louder than the others.
There is also a rights question that has nothing to do with quality. If a campaign uses an ambassador, cloning that voice needs their consent and a licence that covers it. That is a contract conversation, and it is easier before the shoot than after.
Dialect is a casting decision, not a setting
Most pan-regional corporate and institutional work runs in Modern Standard Arabic, because it reads as neutral and travels. Retail and food brands speaking to a specific market usually sound better in the dialect of that market, and the choice between Gulf, Egyptian and Levantine is a positioning decision rather than a technical one. Pick it at script stage. Dubbing a film into the wrong register is a more expensive mistake than any tool saves you.
The hybrid most campaigns end up with
- Record the hero in a booth. The film that carries the brand gets a human read, directed, with the client in the session
- Dub the long tail. Cutdowns, variants and explainers go through AI, reviewed by a native speaker before release
- Subtitle everything. Most feed viewing is silent, so subtitles do more work than either option, in both languages
- Leave timing room in the edit. Arabic copy usually runs longer than English, and a cut built tight to an English read fights every other version
- Keep one master, many layers. Supers, voice and music on separate layers so a new language is an export, not a rebuild
We plan deliverables that way on video production, which is why the LWK + PARTNERS documentary shipped with English, Arabic and Chinese subtitle tracks out of one production block instead of three.
The short version
The split that works in practice: AI for the films people need to understand, a booth for the ones that have to land. Decide which film is which before the edit starts, because retrofitting a language version into a finished cut is where the savings disappear.