Voice actor — tasks, one by one

Narrating audiobooks and e-learning

Long-form narration is the read synthetic voices handle best, because it is long, text-first and tolerant of a steady voice. In May 2025 Audible began offering publishers end-to-end AI narration, with more than 100 AI-generated voices, either as a service it runs or as tools publishers use themselves.

The announcement says what one producer offers, not how many titles now use AI voices or how many narrators lost work; the same company says it keeps investing in premium original content.

Performing characters for animation and games

A lead character is a performance built with a director over many sessions, and audiences notice when it is flat. Synthetic voices are used for placeholder lines and minor characters, and cloned voices can extend an actor's recorded performance — which is why the terms for using a performer's voice this way are now written into contracts.

This rests on the nature of the work, not on a count of productions using synthetic voices for characters; the line between a minor role and a lead is the studio's decision, not the actor's.

Commercials, promos and phone systems

Phone menus, announcements and low-budget explainers are exactly the reads text-to-speech was built for, and they are changed often, which favours a voice that can be regenerated from text. Brand campaigns that trade on a recognisable voice still hire people.

This rests on what the technology can do and on how these reads are commissioned, not on a measurement of how much of this market has moved; the effect is uneven by budget and by language.

Dubbing into other languages

Speech-to-speech translation can now carry a performer's voice into another language, and Audible said in 2025 it would offer this for audiobooks in beta. For screen work, timing to the picture, performance and cultural adaptation are still directed by people, and the new tools mostly change how much of the recording and syncing is done by hand.

This rests on what the tools can do and one company's stated plans, not on a record of how many productions have dubbed this way or how audiences received them.

Deciding who may copy your voice

A voice that can be cloned turns every session contract into a question about future uses. California made part of that question law: a clause letting a digital replica of a performer's voice replace work they would otherwise have done in person is unenforceable, for new performances fixed from 1 January 2025, unless it describes the intended uses reasonably specifically or the performer was represented by a lawyer or by a union whose agreement addresses digital replicas.

This is one state's contract rule; it does not stop digital replicas being made or used, and it says nothing about what performers are paid for them.

Auditions, direction and the home studio

Being cast depends on a person hearing a person and choosing them, and taking direction means changing a read on the spot. Software helps clean audio and manage auditions; it does not get cast.

This rests on how casting works, not on a record of how many roles are now cast without an audition or filled with a synthetic voice from the start.