Why you should be careful with that AI voice buyout

What AI job language looks like

You may see terms such as the following used in an audition breakdown or job contract:

• Text to speech
• Artificial intelligence training
• Machine learning
• Voice doubling (different from “voice matching”)
• Voice modeling
• Simulated voice
• Synthetic voice creation

Why should voice actors be careful about AI?

• Most text-to-speech voices are done for an “in-perpetuity buyout” at a flat rate. The company who purchases the rights to your voice model data can then turn around and sell that to a giant third-party corporation, without any additional compensation to you and perhaps without even your knowledge. But if you already signed your rights away, chances are you will not be able to leverage any further compensation for that usage, no matter how big the third-party company is. For an example of how this can happen, see the TikTok voice lawsuit.
  
  
• While being the (artificially generated) voice of a major brand (think Siri, Alexa, Google Assistant, etc) may sound like a very exciting opportunity to advance your career, it can also backfire on you. If the voice becomes so widely used that the general public associates it with that one specific brand, it may lock you out of potentially lucrative work for other brands—-especially competing ones. There is also no guarantee that you would ever even be allowed to claim credit for being the voice of that brand, as much work of this nature requires contributors to sign very strict NDAs that could even hold them legally liable for disclosing information about the company they worked for.
  
  
• When you sell your voice model data with “no strings attached”, that means anyone who has access to the technology could generate a file of your voice saying whatever they want it to. Some people may have moral objections to their voices being used - even artifically - to promote ideas or brands that don’t align with their personal values. There is also growing concern about “deepfake technology” and how it could be used to harm the reputation of public figures by making it sound like they said something they didn’t.
  
  
• Nowadays, there are certain companies that market the idea of synthetic voices as a cheaper and easier option than hiring actual voice actors. In many industry professionals’ opinion, this type of AI voice usage is the most harmful out of all of them as it can harm the ability of everyone in your field—-including you, your friends, and your peers—-to make a living doing voice work. Please avoid working on this type of job at all costs if you can avoid it, unless they literally want to pay you enough to retire and never have to work in VO for the rest of your life.

AI modeling without consent

While it’s bad enough that actors get talked into willingly selling the rights to their voices for what is a small amount of money in the grand scheme of things, there are sites that scrape data from video games, shows, even YouTube, and brag that you can make your favorite character say anything you want them to. While people who use these sites try to defend it by saying “it’s just for memes, we’re not profiting off of it” and the sites’ terms usually prohibit commercial use, the actors, and their end clients who actually own the copyrights to the recorded voice files, did not consent for their files to be used in this manner. Just like with a lot of AI artwork, these models are cobbled together by feeding stolen assets into an algorithm.

Then, there is also the problem of contract language where AI language is slipped into the contract without the actor noticing. Many contracts tend to be fairly standard, and actors can get used to either signing them without reading, or being afraid to push back for fear of losing the job. It is always important to read over your contracts and make sure you do not accidentally sign your rights away.

If you find your voice being used without your permission, there are various options available to you. If it’s a website that allows users to generate text to speech from your voice (usually as a character you played in a game or show and taken from assets datamined by fans), you may e-mail the owner of the website and request an immediate takedown.

If you have substantial reason to believe that your voice was recreated commercially for a third-party company without your informed consent, you may wish to consult with a lawyer to review your options.


How to protect yourself on non-AI jobs

Whenever you start a new job, or especially when you start working with a new client, check over your contracts to make sure there is no language being used that openly gives them the right to create an AI model of your voice. Terms to watch out for include things like rights to “simulate, synthesize, recreate, double, or model” the actor’s voice. If you come across questionable language, you are within your rights to challenge it or at least ask for clarification. With union jobs, you will have some protection and can contact SAG-AFTRA if you have concerns about clauses in a contract you were asked to sign.

When it comes to nonunion jobs, however, there are no standardized protections, so you may wish to be a little more diligent when it comes to contract terms. Many studio contracts are standardized by their legal teams and cannot be altered, although if you are working with an agent you may be able to get them to push back on any problematic clauses in the contract.




”But the rate they’re offering is really high, and I need the money/work!”

What seems like a large sum of money for a single job actually often turns out to be an insulting offer when you factor in the workload, usage, and potential future ramifications of the job. Even if you take away the idea that your voice could be used for anything in the future without notification or compensation, the workload alone is often more than it seems. For example:

• While some jobs of this nature may be recorded at a studio, a lot of them - especially the lowball ones - expect the actor to record at home. The standards for audio quality tend to be very stringent, so you’ll need to be able to record for hours and hours in a perfectly quiet space with no outside noise.
  
  
• Although a rudimentary voice model can be made without a whole lot of data to go off of, for more sophisticated models, a very large amount of recorded data is required to make the synthetic voice sound as natural as possible. This means you will likely be recording hours and hours of random sentences to feed into the algorithm, leaving you with less time, energy, and vocal stamina for other jobs.
  
  
• In addition to the recording workload alone, chances are you will also be expected to edit your audio to clean out any mistakes, noises, etc. You may think “that doesn’t sound so bad, I already do that for my jobs and auditions anyway!” But cleaning a one-minute audition file is much different than cleaning, say, ten or twenty hours of finished audio. The time adds up, and it probably isn’t going to be compensated extra, as most jobs like this will either pay per finished hour or a flat rate.
  
  
• Clients may also ask for revisions and pickups for any reason they see fit, and these are often unpaid. When a client gets free revisions, they may tend to be excruciatingly nitpicky about the way you pronounced certain words, or what they think is a minor mouth noise or breath, etc. These revisions can be very demanding and likely require a quick turnaround.

Predatory AI voice castings are often posted on pay-to-play casting sites, where clients may hope of suckering in desperate and/or less experienced talent who don’t understand what a job of this nature is worth. The icing on this terrible cake is that the site itself may take a commission fee - perhaps as high as 20-25%! - of your payment in exchange for the use of their platform, making your take-home pay much less than expected.

And of course, when a client is able to artifically generate a recording using your voice any time they wish, it means they don’t have to hire and pay you to actually voice that job for them.

So the next time you see a broad in-perpetuity AI buyout for $3000 and think “wow, that sounds great, I could pay my rent with that!”, know that in context, that rate is insulting and predatory for what they are asking you to do, and it’s not worth giving up the rights to your voice for. It is understandable to want to avoid having a day job, but it is not worth jeopardizing your desired career for a month or two's worth of rent.


If you DO decide to do this type of work

Ultimately it is up to you as an actor to decide whether or not you wish to take a job. If you have been given informed consent about the intended use of your voice for AI and decide for one reason or another to take it, the following tips may help you:

• If it is at all possible to loop your agency in, talk to your agents about the offer before signing anything and ask for their thoughts. Some agents may not want you to do any sort of jobs of this nature, but others may be able to negotiate a better contract for you—-for example, a higher rate, or restrictions on how the voice data can be used. It is well worth the commission to your agent if they can help your rights to be protected.
  
  
• If you are working for one specific company as the artificial voice of their brand, see if you or your agent can negotiate a usage limitation in the contract. For example, let’s say an Internet service provider wants to create a “voice assistant” for their phone lines to help callers diagnose their connection problems. See if you are able to negotiate your voice model to ONLY be used for the purpose of phone systems for that specific company, and ask for it to be spelled out in the contract that they will not sell or give away your voice model data to a third-party company. While clients won’t always agree to it, “single-application” usage protection can help you feel more comfortable about working this type of job.
  
  
• If the nature of the job requires that you absolutely must do a broad all-use buyout that allows them to do anything they want with your synthetic voice model, at the very least, make sure you are compensated extremely well for this. It should cover the fact that it may lock you out of other jobs in the future—-whether jobs you could actually be hired to record, or jobs that you lose by virtue of your voice being associated with a widely-known brand.

Will human voice acting eventually be replaced by AI?

It is too early to know exactly what the future implications will be.

The bad news: As technology continues to evolve, artificial voices are not only sounding more natural and expressive, but consumers are also getting more comfortable interacting with robotic voice assistants in their homes and cars, on their smartphones, and in customer service applications. Not only is AI already creeping into the advertising and narration worlds, unless someone is listening very closely and already familiar with the nuances of a synthetic voice, chances are they won’t even know that voice they heard on that TV commercial was actually read by an AI. Rates were already being driven lower and lower for this type of work, with clients hoping to pay as little as possible, and now with AI they can get voices for even cheaper without having to worry about an actor or their agent trying to negotiate for more money or a contract renewal.

The good news: There will always be a market for the authentic human experience. Even if you can train an AI to sound technically perfect and hit all the right emotional “beats”, it is at its core still soulless. With a human performer, they are bringing all their experiences, emotions, even a part of themselves to each character they play, and audiences tend to deeply connect with those performances. It’s possible that if AI voices go too far, there will be pushback from audiences wanting real people or even companies using the idea of authentic human voices as a selling point. From a marketing perspective, people also like knowing who the voices behind their favorite characters are, going to meet them at conventions and interacting with them on social media. A disembodied voice generated by a computer is simply not the same.