Biden wants a ban on AI voice clones

What happens when voice cloning becomes so convincing that anyone can impersonate a world leader—or your family members—with just a few seconds of audio?

The White House is officially taking notice. President Biden used his State of the Union address on Thursday evening to call for an outright ban on AI voice impersonations.

While outlining his unity agenda and priorities for a potential second term, Biden grouped the technology alongside top national concerns. He urged lawmakers to strengthen penalties on fentanyl trafficking, pass bipartisan child privacy laws, harness AI safely, and ban AI voice impersonations.

A big callout with zero specifics

The president did not elaborate on how this ban would actually work.

He offered no details regarding the types of guardrails or penalties he plans to push, nor did he clarify whether the restriction would extend into the entertainment sector.

Voice replication has already been ground zero for labor battles. Last year, SAG-AFTRA made AI protection a centerpiece of its prolonged strike against Hollywood studios.

That dispute ended with landmark guardrails requiring studios to secure explicit consent and provide compensation before replicating an actor's voice or likeness. Yet, that deal left a massive loophole: it did not stop studios from training AI models to produce completely synthetic performers who do not resemble real people.

Biden’s high-stakes speech also featured moments of political theater. Broadcast cameras caught Georgia Republican Rep. Marjorie Taylor Greene wearing a red MAGA hat and heckling the president as he spoke.

Why it matters

Voice cloning tech has gotten insanely good and dangerously accessible, shifting from viral internet novelties to sophisticated phone scams and political disinformation almost overnight.

For regular consumers and creators, federal recognition of voice cloning risks is welcome news. Your voice is personal property, and the current legal system barely knows how to protect it. But the real test lies in execution: outright bans sound great on a podium, yet drawing the line between malicious deepfakes, fair-use satire, and legit Hollywood dubbing is a legal nightmare.

For AI audio startups and big tech, this is an unmistakable warning shot. Companies building voice generation models will likely face tighter identity verification requirements, watermarking mandates, and audit trails sooner rather than later.

We think dropping this demand into the State of the Union makes for great politics, but treating voice cloning with a simple ban might be unrealistic. Until the administration lays out concrete legislative language, this looks more like an opening salvo than a real policy framework.