These are perilous times in the games industry. Seemingly every week, workers are being laid off en masse, studios are closing, and unannounced games are being canceled. Even the giants of the industry are not immune from the bloodletting. And hanging over it all is the spectre (or promise, depending on your perspective) of Artificial Intelligence. Ever since AI tools, such as MidJourney and ChatGPT, hit the mainstream, the topic has been at the forefront of public consciousness and discussion, and even more so in the gaming space.
Those in the industry have always used “artificial intelligence” in reference to the “human-ness” and smarts with which enemies and NPC’s act, but now it has taken on an altogether different meaning as entire games can be created with AI-generated assets. One of the facets of game development seemingly most at risk due to the explosive development of AI is voice acting. Already, games are coming to market that feature significant portions of their voiced dialogue generated by AI. Following the recent SAG-AFTRA strike, the union made a deal with an AI company to license actors’ voices for use in games, leaving many video game Voice Actors angry with a deal they said they had no input on. It all leads to a fraught situation with many questioning the future of the industry.
It is in this light that we came to interview Ryan Cooper, an accomplished voice actor across indie games, audiobooks, and more. In our recent review of the indie Metroidvania The Last Faith, in which Ryan voiced The Old Wymond, among other characters, we specifically praised the quality of the writing and voice acting. Ryan reached out to us, and after I spent some time looking at an experiment he did on his blog that compared his own work with the results that AI-generation tools could produce, I asked if we could put together a full interview. Thus, we now present this conversation, with topics ranging from his start in the industry to performing with live direction, the perils of AI voice acting, and much more!
Seasoned Gaming: I always like to start by giving our audience an understanding of how our guests arrived at this moment, about your path into the gaming industry, and how you came to be a voice actor. It’s so different for everyone, and there’s no one way or right way to do it, so what did that journey look like for you?
As a boy living in our semi-detached house in the West Midlands of England, I remember marvelling over a simple tape deck recorder that allowed me to actually listen back to my own voice. My brothers and I would play pretend DJ for mixtape interstitials. We would later upgrade to furtive communications over ham radio: “Rubber duck, looks like we got ourselves a convoy.” There was a thrill in listening and speaking in all these whacky voices. I think the foundations were laid then.
My route to voice acting in earnest began with a serious knee injury sustained during another passion of mine, Soccer. I had enough downtime to reach the point of boredom required for true personal reinvention. I caught wind of a fan project called Skywind, and they were after voice actors; my childlike curiosity piqued again. I get to pretend to be someone else, record it, and do it in homage to one my favourite games of all time, Morrowind? Where do I sign up?
In the years since, it has been a case of accruing the knowledge, the setup, the experience and learning how to put myself out into the world. I was fortunate to get in early with three terrific projects that gave me the belief I could do it, namely The Old City: Leviathan, Stasis and Brigador.
Seasoned Gaming: When you work on a game project, is it a situation where you simply get the script and record your lines in a vacuum, or is there a lot of back-and-forth collaboration with the developers and writers of the game?
Some clients have a clear vision for a role even down to specific references; others prefer to put you in the driver’s seat. I will record a sample for feedback, and gradually zero in on it. That’s the back-and-forth collaboration you mention. Sometimes you will be asked to “self-direct,” others you will be live directed by the client at the other end of a Discord call, in which the whole thing is wrapped up in one or two sittings. Self-direction and live direction are different approaches with their own pros and cons.
Seasoned Gaming: Will you typically record at home or do you go into a studio for that? You have quite a few credits in 2020 and 2021, how did the pandemic affect your work?
My experience has almost exclusively been from my private studio. I am more than happy to travel for the right project, but where I live now in the Southwest is a quiet part of England and opportunities are rare with local studios. The advantage of online voice acting is my availability to the entire world. I have worked with clients in North America, South America, South Africa, Australia, and throughout mainland Europe, which would be impossible had they insisted I pop in to see them.
Regarding the pandemic: you do get quick turnarounds, but indie development cycles are typically several years, and there is no fixed point at which voice acting is tackled. It’s possible projects that came out in 2020 were voiced circa 2018. I was first attached to Ghost Song way back in 2013! So, it’s likely my studio work during the worst of the pandemic was in projects slated for 2022, 2023. My vague recollection is there was something of a dip, but my fears were more for my day job. Thankfully, we pulled through.
Seasoned Gaming: What is your creative process, if you are willing to share a bit of that? How do you prepare, understand the character, and get into the mindset for the performance?
It is exciting, and daunting, to start with nothing. I go deeper into my process on my website but to paraphrase: voice acting is a cerebral artform that requires a period of forethought. A complete read of the script goes without saying. I like to reflect on the themes, the characters, the milieu and where my guy fits into it all. Then, I break him down to basics. Keywords help: Deceit. This guy is hiding an ulterior motive. Savant. He is a science genius – who hasn’t the first clue about people. Vulnerable. He’s alone in a hostile place searching for his family.
Having a working knowledge of popular culture is reassuring. Okay, this guy gives me late-stage Richard Harris vibes, with a splash of Anthony Hopkins. Now, I no longer have nothing, I have a springboard. If you spend enough time thinking, the answers come to you in unexpected moments. In the car, walking the dog… trying to get to sleep. I’ve had to assure people that I am not some madman after they heard me mumbling to no one under my breath. I am at the point where my thoughts are evolving into practice, and it can overspill a little.
You build a character, slip inside and lock yourself in. It follows you around for weeks, and fades after you’re done with it. It’s not a tap you can turn on and off at will, so I understand actors’ frustration when they are asked by talk show hosts to ‘do that voice’. I reprised a role several years later and was surprised by how difficult it was. I had to go through the process again. Ah yes, I remember now, I made that sound by speaking up through the nasal cavity, I trapped it in that particular part of the throat…
Seasoned Gaming: Now that we have a background on what you do, I wanted to get into the topic that is consuming a lot of the oxygen in the industry these days: the use of AI in games for voice work. You have a post on your blog from almost a year ago, now, that compared some of your work with what an AI creation tool produced when you “trained” it on those same performances. What came of that was really fascinating to hear. I know from that blog post and your comments to me via email that you have a lot of thoughts on the subject, so I’d like to just give you the floor and let you take us where you will on the topic.
With the current focus on voice acting, it would be remiss not to mention this affects all facets of the industry. Writers, artists, asset creators, animators, sound designers, all are prospectively under threat. We assumed AI would do the rote, analytical stuff while never coming close to creativity. How little we knew…
There is something particularly unsettling about a human voice produced by a machine. It’s the uncanny valley for the ears. More unsettling is the dreadful practice of cloning actor’s voices without their knowledge (let alone their consent) to generate transgressive content. It is not that AI itself is immoral, it’s that it is being wielded by immoral humans for immoral purposes. I sympathise with SAG-AFTRA members who feel betrayed by the union for pallying up with an AI voice firm, precisely the thing they were striking against.
The kicker is: AI voice acting isn’t entirely unethical. For instance, Miłogost Reczek (the voice actor for Viktor Vektor in Cyberpunk 2077) died after the game’s release, but before the Phantom Liberty DLC. With his family’s permission, CD Projeckt Red reconstructed Reczek’s voice using AI, which they achieved by overlaying it onto the performance of another VA. Everything was handled respectfully, involved a paid human actor and avoided undesirable alternatives, like omitting Viktor from the story entirely or recasting him. I’m cool with that.
There is another instance in which it’s hard to make an ethical case against AI voice acting. Let’s say a small indie studio has dreams of a sprawling story-driven magnum opus, in which the player will engage with hundreds of NPCs. Only, they have an indie budget scrounged from personal savings or a modest Kickstarter success, and can only to afford to hire a core group of ten voice actors for the main leads and supporting roles. Is it okay to give passable AI voices to the otherwise unvoiced minor NPCs, to fulfil the original vision and maintain immersion? Is using AI in this way less ethical than hiring unpaid actors you found on Casting Call Club, while you go on to enjoy the fruits of their labour?
However, even if we accept with caveats that AI is fair here, this scenario quickly becomes dystopic. What if the developer reconsiders their financial situation and concludes it would be a huge cost saving and time saving to replace the ten voice actors as well? Sure, any semblance of humanity is stripped from the performances… but I do get my game out on Steam ahead of schedule and under budget. If indie devs must weigh this up in the years to come, you can bet the AAA guys will too – only it will pose far less of an ethical quandary to them. There are inherently corporate implications to AI voice acting: get the product out quicker and cheaper. Maximise the profit margins!
Seasoned Gaming: It’s widely discussed how fast AI has improved in the last year, have you gone back since that blog post and recreated the exercise to see if the AI performances have changed or improved?
I haven’t, mostly because the Eleven Labs experiment was a free trial, and I am far too much of a tightwad to spend money on it! AI has irrefutably improved in a short time, but I am yet to be convinced. There was a recent controversy involving a YouTuber who digitally resurrected legendary stand-up comic George Carlin to perform a one-hour special about modern life. Predictably it caused outrage, not least from Carlin’s estate (who I believe are suing). Against my better judgement I had a listen, just to check on the competition. My initial response was, “Oh, crap, it’s over.” But the more I listened, I began to relax.
It’s Carlin if he were hiding a deep depression from the audience, or perhaps the back end of a cold. There’s no pop. It is listless in a way that is intuitively not so with his stand-up. Dum-de-dum-de-dum-de-dum [fake applause], rinse and repeat. The impish qualities of Carlin are removed, like the way he brought his voice almost to a hush to reel you in, and then went big with the punchline. The subconscious cues only a human gets, the way his body language up on stage informed his inflections – not possible for an algorithm. The vital human spark is missing.
You may have noticed there are no examples of AI voices raging, pleading, holding back tears, or giving in to unbridled joy. Why, pray tell, is that? Without emotion it is not acting, it’s a recital. Only when it breaks that ceiling will I worry. Indeed, voice actors being put out of work would be the least of our concerns next to the more existential fears that arise from a perfect simulacrum of all human emotion and experience. You have deeply philosophical arguments about sentience to contend with at that point.
What worries me more is if you marinade people in such an artifice for long enough (like with CGI imagery), that this intuition will atrophy. 75% real will be real enough. I can see certain audiobook readers being put out of work. Brigador and Brigador Killers are impassioned works of which I am sceptical an AI could ever match. On the other hand, Dungeons of Infinity or the Skyblivion developer diaries, are more equanimous; an AI could well match those ones. The dread comes from if / how AI will shrink the work supply while the talent pool remains the same, creating a brutally competitive market among humans. I hold this fear for all human jobs. AI is not just a repeat of the industrial revolution; it is something unprecedented that may well stymy all human potential.
Seasoned Gaming: I’m intrigued by your discussion of the grey areas in the “AI Bad!” narrative, and it brought to mind another situation I’d like to get your thoughts on. First, if you’ve had a chance to listen to the AI commentators in the multiplayer FPS “The Finals,” I’m wondering what you might think of those performances.
[Goes away and listens to The Finals]
Frankly, they suck. The voices are more declarative, but the same congenital issues are there. There’s hardly a line that is enunciated the way you would expect, with odd cadences and emphasis on the wrong syllables. “The KingFISH!” “One contestant ISTRYINGTOKEEPTHEMINTHEGAME!!!” As a character-based actor whose bread-and-butter is in motivations, emotions, arcs and relationships, I don’t feel threatened by this. I do expect it to improve, however, so we may yet add, “cheesy game show host” to the endangered species list along with audiobook narrator.
Seasoned Gaming: Extending that idea, what do you think of a scenario where AI might be able to make a game more immersive, where narration could comment directly on what a player does in an open-world sandbox game, or commentary in sports games could potentially be more engrossing and closer to emulating the live TV experience?
We’re talking about a game where, not only are the lines performed artificially, they are also adaptive and generate on the fly in response to the player’s decisions. I too have daydreamed about such a game. It would represent an awesome leap, as it would essentially mean an AI ‘lives’ inside your game that can critically analyse your performance and formulate lifelike responses. Maybe even anticipate what you will do next.
Even if it were an ‘extension’ of a human performance, once more troubling questions about sentience – even topics like the Simulation hypothesis, are raised. Is each copy of the AI learning more about human sport (or how humans go about killing one another) as it observes? Is it reporting back to a Mother AI via your Internet connection? I shudder to think. If the alternative is that stuffy old fella giving me the same played-out canned responses that he did in Cash Grab Sport Sim ‘23, I would prefer that!
Seasoned Gaming: On a related topic, 2024 has already seen an unprecedented number of layoffs at studios and publishers of all sizes, across all sorts of roles. Voice actors like yourselves are not typically employed by these entities, so how do you see these happenings affecting your part of the industry? Do you think the layoffs are just the logical result of consolidation and the difficulties of forecasting the future during the pandemic, or is there something more endemic and troubling at play? What would you do to fix the problem if you could just wave a magic wand and have it happen?
That’s one Hell of a question, one that probably escalates as high as geopolitics and global economics. It seems since the pandemic, we have lurched from one global disaster to the next. We’ve suffered runaway inflation and the cost-of-living crisis, the wars in Ukraine and Gaza, various supply line issues and a general uptick in world strife. All impact our ability to live and how much disposable income we have. It’s not pleasant to admit, but entertainment is a luxury compared to food, housing and health. So, to wave my magic wand, finger-snap my Infinity Gauntlet or thrice click my heels, I would will things back to the time before when things were reassuringly dull.
I’d like to think I am somewhat resistant to these global forces insomuch that I am small – small meaning nimble and adaptable. I do not operate in the sphere of an EA or an Activision Blizzard; I freelance for companies whose employees number closer to ten than ten thousand. Smaller entities are less likely to feel the pain of sudden market shocks. Cutting staff by 6% may not seem that bad on the face of it, but that’s hundreds of affected families in the case of AAA.
The ~6000 staff who have lost their jobs in 2024 alone have my deepest sympathies. There are worse things, however, than being in a position to form a new indie studio, over which you have creative control. If you need a voice actor, I might know someone… /shameless plug
Seasoned Gaming: You have a fair number of audiobooks on your CV; how did you find your way to that type of VA, and how does your process with those roles differ from that for games?
My more recent audiobooks are with Forteller, who provide accompanying audio narration for board games. They typically put out public casting calls, so it’s about impressing them with your range. Brigador was different. I believe I approached Stellar Jockeys, in the hope of voicing a character for what I presumed would be the game’s story campaign. To my surprise, they asked me to voice an accompanying 80,000-word audiobook instead (the first of its kind on Steam).
If a character performance is a 100m sprint, an audiobook is a marathon. For the sequel Brigador Killers, I was live directed by a lovely gent called Benjamin Glover for one hour, three days a week, for five months solid. This thing became a part of my life and squatted in my brain for a long time. I was screaming obscenities at the height of my lungs, praying the neighbours weren’t dialling the Police. I finished up with that strange melancholy you get when you finish a good book, only way more intense. The process for the original Brigador was much the opposite; I was self-directed, while starting a new job. I was squeezing in morning sessions before I drove off and paying my brother Glen a percentage to chop up the rushes. Less polished, but certainly raw.
I interpreted the narrator as a character of that world in his own right, and then played a good ten characters on top of that. That’s a challenge. You cannot go through the process in the same depth, so you ascribe little markers like a rasp, an odd cadence, or the use of a head voice opposite a chest voice to make a distinction. I think people could sense which character was which, even though I was essentially speaking in the same voice.
Seasoned Gaming: Would you say that working with a live director was more difficult than recording on your own? Do you think it resulted in the performance being different or better than your self-directed (for lack of a better term) work?
Killers was a comfortable experience once I found my groove, and as I got to know Benjamin better. I had already developed the tone for the Brigador narrator years before, so that was like an old glove. Benjamin filled me in on the new characters, down to extratextual details like, “he was a smoker for decades”. Moreover, we placed an emphasis on this inexorable, accelerating intensity, and dread. It is the archetypal dark sequel centred on the theme of consequences.
I find live direction to be more challenging, but that could be because relatively little of my experience has come out of it. Others might thrive on the pressure of a live performance with an audience, and balk at the additional requirements of self-direction, whereas I am in my element. There is no self-consciousness, no one listening… or judging. There are the demands of a deadline sure, but you can work through it as and when you are in the mood.
Another challenge is how much of “the process” one should commit to. You can over–prepare. For one session, I spent hours studying certain influences, speech patterns etc. I tried it out and immediately the director was like, “Yeah, we’re not doing that.” You can become unmoored. There is a temptation to play it safe and not add many of what I call “modifiers” to the voice, lest you cannot replicate them on demand.
I remember watching a video of a live-directed AAA production. To paraphrase, it was, “Thanks. Now do a take exactly like that, but a couple of octaves lower… Hmm, too much, go up a little. Great. Now hold that, but speed it up 15% and put a little rasp in the back end. No, not too much. Okay, now one with…” It was less the use of context to subtly guide the actor’s performance, and more micromanaging his every utterance. It was so technical you might as well have had an AI doing it. Thankfully that has not been my experience, and I respect that directors have their own challenges and pressures to contend with.
All that said, there are definite advantages to live direction. The obvious one is it’s less onerous. Self-direction is three roles: actor, director and basic QA guy, before you can deliver to the client. Can’t let the flubs, farts and swearwords make the cut, can we… “This line is bloody terrible! Who wrote this?” Because I am a finicky git, I can’t review my work with monitoring headphones like a normal person. It feels too… close, and not how I imagine most gamers experience the sound. I need to listen to it ‘in the air’. This all adds time and inexpedience, and because you agreed a flat rate in advance (completed word or line), it’s probably not time for which you will be compensated. Self-directed and live directed sessions have different payment structures, the latter paying better pound-for-pound.
Seasoned Gaming: What have been some of your favorite parts to play over the course of your career?
There have been a few this past decade! John Maracheck (the protagonist of Stasis) is perhaps the performance for which I am best known; it is my rawest and most emotive role. I was pleased to learn Mandalore Gaming had praised my work on the game recently. Then there is Boleslav, the first King of Poland whom I portrayed in Ancestors Legacy. It was nailed in a single two-hour session, and I was out on my feet, drenched in sweat by the end of it. The guy is just unrelenting aggression. It was also an attempted Polish accent (in the presence of a Polish live director, no less), which was quite the challenge.
I also grew fond of Crain Tallengyr for the Frosthaven audio narration. He is a plucky, eccentric engineer whose plans almost always lead his colleagues to near-death experiences. Lately, I have found I am taking on smaller but more numerous parts within the same project. Between the Metroidvanias Ghost Song and The Last Faith, I produced eleven character performances which was real fun. I particularly like The Old Wymond as his is a voice that departs far from my conventional speaking range. There is an extra layer to him that I don’t think fans have sussed out yet.
Seasoned Gaming: Are there any game franchises you’d particularly like to work on, or any established characters you’d love to play if given the chance?
I have been a huge Batman fan since I was a kid. It’s more like three roles: the Batman, the playboy and the more private Bruce Wayne – all quite distinct. The complexities, the contradictions, the dissonant way parts of him mirror his adversaries, the instantly recognisable trauma at the heart of the role, it’s legendary at this point. Batman has remained in the collective psyche for over 80 years because he can be reinterpreted, seemingly without limit. “He’s my Batman”. Everyone possesses their own ideal version of the character.
As for franchises, The Elder Scrolls is the one I would love to work on the most. My voicework features in some of the biggest Skyrim mods: Skywind, Skyblivion, Beyond Skyrim, Enderal etc. I have also collaborated with prominent Elder Scrolls YouTuber Camelworks and leant my voice to the Fallout 4 add-on Sim Settlements.
Seasoned Gaming: What are your go-to game genres and what are you playing these days?
I love a good RPG, or anything that immerses me in a world. Did I mention I love Batman? Yeah, I love the Arkham games too. Ever since the days of Super Metroid and Castlevania IV, I have always had a soft spot for Metroidvanias and take pride in voicing for a couple of well-received examples in recent times. I have actually honed my hand-eye coordination as to play games while walking on a treadmill, so I can exercise and play at the same time. I am down to only one spectacular fall per year!
I’d like to give a huge thanks to Ryan Cooper for jumping into all my questions with gusto and producing some fantastic, thought-provoking discussions. I highly encourage our readers to make their way over to his blog and listen to the comparisons of his performances with their AI analogues. It’s illuminating in every way and I’m both fascinated and terrified by where the gaming industry will end up in the era of AI.
