Meta is offering Hollywood stars millions for AI voice rights

Meta is opening its checkbook to bring Hollywood royalty into your ears.

According to reports from Bloomberg and The New York Times, Meta is currently in negotiations with several top Hollywood agencies, offering millions of dollars to license celebrity voices for its AI products.

Among the names on the table are Judi Dench, Awkwafina, and Keegan-Michael Key.

Hollywood contracts and high-stakes talks

Meta wants to take recorded audio from these stars and integrate them across its Meta AI digital assistant. The rollout would span Facebook, Instagram, Messenger, WhatsApp, and the Ray-Ban Meta smart glasses.

The company is reportedly rushing to finalize deals before its annual Connect conference in September, where it plans to unveil a fresh suite of AI features.

Getting there has not been a walk in the park. Talks have started and stopped multiple times due to disagreements over how the voices can be used. For now, the parties seem to have settled on strict time limits: Meta would only own the rights for a limited period, with actors holding the option to renew or walk away.

Interestingly, the SAG-AFTRA actors union has reportedly reached an agreement with Meta on baseline terms. Following last year's historic strike, the union established strict rules requiring consent and compensation before using an actor's digital likeness, and Meta appears to be playing by the book.

Moving past last year's chatbot graveyard

If this strategy sounds familiar, that is because Meta tried a version of this just last year.

At Connect 2023, the company rolled out 28 celebrity-themed AI text personas modeled after figures like Snoop Dogg, Paris Hilton, and Kendall Jenner. But according to a recent report from The Information, Meta quietly pulled the plug on those chatbots, and their profiles have vanished.

Text chatbots mimicking celebrities felt like an ephemeral gimmick. But real-time, expressive speech synthesis is an entirely different beast. Ever since OpenAI demonstrated GPT-4o's human-like vocal performance, the race has been on to make conversational AI sound genuinely engaging rather than robotic.

Why it matters

We think this move is clearly aimed at driving adoption for the Ray-Ban Meta glasses. Audio-first wearable interfaces live or die by vocal quality. Hearing an iconic, expressive voice in your ear instead of a flat synthetic readout turns an awkward gadget into a compelling daily companion.

From an industry standpoint, Meta is taking a direct swing at OpenAI's recent copyright headaches. After OpenAI faced intense backlash over a GPT-4o voice that sounded suspiciously like Scarlett Johansson, Meta is taking the expensive, legal route through SAG-AFTRA. That sets a strong precedent that frontier voice agents require formal studio contracts, forcing competitors like Apple and Google to rethink their licensing playbook.

The real risk here is novelty fatigue. Meta already learned the hard way that famous faces cannot save a mediocre text tool. If the underlying Meta AI cannot provide meaningful daily utility, paying millions for Hollywood vocal cords will simply amount to the world's most expensive novelty ringtone.