The voice behind the machine: How AI is changing voice acting
The growing impact of AI is reshaping how the voice-over industry operates
In the early 2020s, the rise of ChatGPT and other AI platforms transformed how many activities are performed, making processes faster, more efficient and scalable. However, this technological shift has also increased concerns about AI replacing human professionals.
ًTechnology becoming a threat for the entertainment industry
When it comes to the dubbing industry, the demand for it is still very relevant, and the tendency is to keep growing. With the rise of streaming services, which allow media content (TV shows, soap operas, movies, documentaries, etc.) to reach people all over the world, global audiovisual content isn’t so niche and locally distributed anymore.
The main streaming services like Netflix and Amazon Prime Video are investing in adding more language options for their content so they can increase their share of the international markets. For example, there’s an estimate that 43% of the viewers in countries like Italy, Germany, France and the UK choose to watch dubbed content over subtitled versions.
Netflix’s co-CEO Ted Sarandos, in 2025, stated that he’s been testing on new TV shows the usage of AI for various purposes, one of them being a way for AI to synchronise the character’s lip movements with the dubbing to improve realism and the viewer experience. On the other side, Prime Video created an AI-aided dubbed system supporting some series and movies to generate automatically dubbing in English and Spanish.
How AI is changing voice acting in E-Learning, audiobooks and video games
The risk of voice-over acting being replaced by AI follows a spectrum, likely to occur where work is highly repetitive, standardised, and produced at high volume. Each type of content therefore requires a different analysis.
For instance, IVR systems, e-learning, corporate explainers and low-budget online content are particularly exposed because they often prioritise clarity, budget, consistency and speed over complex acting and performance.
To understand the actual harm from the perspective of the professionals in the area, we reached out to Carla Gmurczyk, a 35-year-old Brazilian who has been a voice actress for the past 15 years. She shared that:
‘Many professional voice actors have lost recurring contracts, such a as long e-learning recordings or IVR phone menus, because the companies decided to switch to synthetic voice banks, which allows instant script updates with no studio costs.’
In the middle of the spectrum are audiobooks and commercials where hybrid models may become increasingly common: AI can handle some of the low-cost or repetitive material, while human actors remain valuable for projects requiring stronger emotional interpretation and appeal, character development and also a recognisable voice. In a 2026 experiment with listeners of literary audiobooks, Ahrens-Schwabe and colleagues found only minimal differences in how people rated AI-generated and professional human narration and concluded that, from the listener's perspective, AI voices can currently serve as a substitute.
Animation and video games occupy a more complicated position. AI can be useful for prototypes, placeholder tracks, background characters and repeated lines, while major roles still depend heavily on human performers and their ability to create distinctive characters and emotional performances.
Whether AI will eventually put significant pressure on the most performative areas, such as lead game roles and major commercials, remains an open question.
The legal battle: Copyright, contracts and consent
The legal debate over voice cloning is developing across four main areas: copyright, personality rights, contracts and regulation. Together, they reveal a central difficulty: existing laws do not always clearly answer who controls an AI-generated version of a person’s voice.
For starters, copyright doesn’t necessarily protect a person’s voice itself. Copying a recording and recreating a voice's characteristics are not necessarily the same legal act. An AI system can generate a new recording that sounds like a particular person without reproducing the original, so a voice can be highly identifiable and commercially valuable without being a copyrighted work.
Contracts and consent add another layer of protection and uncertainty. Their terms can determine what a performer has authorised, including whether recordings may be used for AI training, cloning or future commercial uses.
Get 5AM Authentic Marketing’s stories in your inbox
Join Medium for free to get updates from this writer.
The argument between Scarlett Johansson and OpenAI is a good example that gained exposure across the industry. The actress says Sam Altman, the company's CEO, asked her to voice over ChatGPT, which she declined. When OpenAI introduced its voice called “Sky”, it sounded so similar to hers that even her closest friends couldn’t tell the difference. Altman had also posted the word “her” around the launch, which she saw as a reference to the film in which she performed as an AI assistant.
This brought questions of consent, imitation and personality rights into public debate, and also showed how hard the legal question becomes when a voice is not literally copied but is perceived as belonging to a recognisable person.
Authenticity and emotion: Does the audience care whether a voice is performed by a human?
The crucial part about AI voice-overs is the fact that emotion, interpretation, and direction matter and timing can’t be robotised. There are aspects in acting and narration that are irreplaceable, and only a human being can translate them into words and sounds.
Amazon Prime Video's AI-aided dubbing systems aimed to make its shows more accessible, but there’s a silver lining: audiences didn’t necessarily react as expected. For the anime fans, some viewers mentioned that the dubbing for shows like “Banana Fish,” and “No Game, No Life“ are emotionless, and the lack of interpretation is making their voices strange and feel disconnected from the characters on screen, affecting their experience
In the interview with Carla Gmurczyk she highlights that:
‘Projects that demand deep sensitivity, artistic direction, complex dubbing, and strong emotional connection, human performance remains irreplaceable.’
The index above shows consumers' perceptions of AI-generated voices in advertisements. As many as 36% of the participants believe that the synthetic voice makes the ad feel antithetical or manipulative. It also shows that only 13% of consumers are likely to buy a product when the ad is made by AI; 41% are bothered by these ads, and Gen Z is less likely to buy a product marketed in this format.
The future of the profession
According to Fabio Azevedo, a famous voice-over actor in Brazil and president of the Brazilian Association of Dubbing Professionals, more than 2 million full-time and part-time voice actors worldwide face the possibility of losing their livelihood and the rights to their own voice. The voice actors even mentioned that they are not only losing job opportunities, but that their voices have been used without authorisation for the training of AI models in order to make them more realistic, without proper payment or recognition, and, most of the time, without even knowing it.
In 2023, a massive strike happened in Hollywood, interrupting numerous shows and movies worldwide. One of the goals of the strike was to discuss a way of protecting actors, translators, writers and voice actors from the threat of the rise of AI and its unfair use. Many other countries are also in the fight to develop their own laws and regulations in order to protect their own professionals, creating limits for unauthorised AI.
The industry is also measuring the trend. In the National Association of Voice Actors’ 2026 State of Voiceover survey, 21% of the participants said they had knowingly lost work to synthetic voices, up from 14% a year earlier.
Carla agrees that AI is already a threat to the actor´s livelihood and that the demand for projects are continuously dropping while existing contracts are not being renewed. When asked if there is a way for AI to be part of the process without harming the professionals, she points out this technological advances could be used to accelerate editing and cleanup process, so the workers within the voice-over industry could focus their time solely on performing and increasing the number of projects they could be part of.
The whole discussion about the future of the profession is coming to a final answer: voice acting will not simply end. Yes, the technology will most likely replace some of the voices and dubbing with AI generated voices, but real sentiment can’t be artificially created. Being able to really act and incorporate the emotions will be way more valuable than just having a lovely voice.
The aim of the discussion isn’t about stopping the advances of AI in audiovisual, but it is to regulate and protect the professionals involved, making sure everyone’s rights are guaranteed and fairly compensated. The argument is also an invitation to discuss the changes in the industry in recent years related to AI, but also how this same industry has already neglected these voice actors and writers for years, and now the world finally gets to see the faces behind those voices outside of the big screens.
Have you ever heard of this issue before?
If you are a voice actor or writer and experienced this growing issue, please share your story with us!
Follow us on Medium 5AM Authentic Marketing and go check our story about programs to help international students living in France.
About this article
This article has been written by a student on the Grenoble Ecole de Management’s Advanced Masters in Digital Strategy Management. As part of a content creation assignment, students are given the task of writing articles based on their digital interests and disseminating the articles online. Articles are marked but we make minimal changes to the content. Thanks for reading! James Barisic, Programme Director, MS DSM.
