The AI Voice Revolution is Bigger Than ScarJo

Global film and TV, libraries and costs are about to be upended by sophisticated speech models and lip-syncing tech. Ready for 'Oppenheimer' in perfect Italian?

THAT’S HER Scarlett Johansson waved to the crowd at the Her premiere in Rome in 2013, a decade before her dystopian portrayal of an AI companion became reality. (Vittorio Zunino Celotto/Getty Images)

If you’re reading this, you hopefully care about the intersection of AI and Hollywood. The recent controversy between OpenAI and Scarlett Johansson and whether the tech company used her voice for its new conversational chatbot without permission is almost certainly one of the biggest AI Hollywood stories to date. Even if OpenAI CEO Sam Altman merely wanted everyone to think that the Her star was the voice of the AI bot — as his tweet the day of the launch suggests — well, you get international news.

All the debates about persona rights and cloning voices are fascinating, but there’s actually a bigger story that hasn’t been fully explored (apologies to both ScarJo and her legal team) and that’s how AI voice dubbing and lip-syncing are becoming ever more popular, accessible, high-quality — and increasingly transformative to international TV markets.

There are a ton of use cases for AI voice replications, but for me, as an international producer, the place I see it being most useful — and most disruptive — is in localization. 

For a long time, comically poor dubs of Bruce Lee films created a stereotype about the practice. But localization has grown into a whole business in itself, and it's incredibly important for international sales deals. Subtitles are fine, but most people — regardless of culture, society, location — prefer dubbing if it’s available. There are, of course, exceptions (think Shogun), but generally, dubbing is the best way to get your content seen by more people.

With AI, the world will have access to more content from more cultures than ever before. AI is going to break language barriers.

Today, less than one percent of content gets dubbed into another language. Think about how the world will change if we increase that even to five or 10 percent. Global libraries could expand across country borders 10-fold overnight. The problem up until now, of course, relates to costs. Dubs are often 15-20 times the cost of subtitles. What happens if AI brings these costs on par? What happens to an entire industry that is suddenly operating at a fraction of its current budgets?

And what does it mean for all of the people working and making a living from traditional dubbing?

In this article, you’ll learn:

• Why the traditional dubbing model is limited by a “constant cost-benefit analysis”

• The sophisticated AI dubbing tool that could completely transform global TV and film 

• How one AI dubbing tool is prioritizing human involvement — and actor consent and compensation

• The primary, unsettled legal issues surrounding AI voices

• Why new AI lip-dubbing technology can further mainstream global dubbing

• How these AI tools vastly expand the market for — and value of — library content

• What SAG is doing to get actors paid and maintain control of their voices

• The unresolved ethical issues that we as an industry still need to resolve