A prospective client recently asked me if my performance could compete with an AI-generated read.
That’s not a hypothetical anymore. That’s a Tuesday. Voice seekers are continually weighing options of using human vs. AI-generated voices.
I didn’t get defensive about it, and I’m not going to get defensive here either. I’ve spent the last couple of years going deeper into AI than most voice actors bother to — prompt engineering, tools, workflows, the whole thing. I’ve run my own scripts through generative voice tools and listened back against my own reads, side by side, on the same copy. So when I tell you what I think AI voice tech can’t do yet, I’m not speaking from fear of it. I’m speaking from having actually used it, tested it, and heard exactly where it holds up and where it doesn’t.
This post is about where that line really sits. Not the scary version. The honest one — the kind of straight answer I try to give on every project that comes through Archbold Media Services.
I’m Not Here to Trash AI — I’ve Been in the Room With It
Here’s something most of the “AI is coming for voice actors” posts won’t tell you: the person writing them usually hasn’t spent real time with the tools.
I have. I completed a prompt engineering specialization through Vanderbilt, and I’ve spent time building small AI-powered tools and systems for my own business — not as a hobby, but because I think understanding this stuff from the inside is part of staying relevant. Sales taught me how to read a room. On-camera acting taught me how to read a moment. AI is teaching me how to read a system.
That matters here, because it means I’m not writing this from a place of “protect the old way of doing things.” I’m writing it from a place of “I’ve seen what the new way of doing things is actually good at, and where it still runs out of road.” It’s the same instinct that shaped what I learned moving from a sales career into booking voiceover work — understand the thing fully before you decide what it’s actually worth.
You can’t credibly defend the value of human performance if you’ve never taken the tools seriously enough to test them.
What Generative AI Voice Tech Actually Does Well
Let’s give credit where it’s due, because pretending AI voice technology is bad just isn’t accurate anymore — it’s come a long way, and dismissing it outright ignores that.
Generative AI is genuinely strong at:
- A rough draft narration in minutes, not days.
- The same tone, pace, and pronunciation across hundreds of lines without fatigue.
- Budget flexibility. For high-volume, low-stakes content — internal training modules, placeholder scratch tracks, rapid prototyping — it’s often the smarter business decision.
- Scripts change late? An AI voice actor doesn’t need to be re-booked.
If you’re a production company or an eLearning team weighing options, none of that should be dismissed. There are real use cases where AI voice technology is the right tool for the job.
But “the right tool for some jobs” and “a replacement for the craft” are two very different claims. That’s where I want to spend the rest of this post.
Where the Line Actually Sits — Intention, Relationship, and Listening
Here’s what I’ve noticed, both from testing AI tools myself and from twenty-plus years of reading rooms in sales and on set: the gap isn’t really about sound. AI voice technology can sound remarkably human now. The gap is about judgment — and judgment is what actually determines whether a performance lands with an audience or just washes over them. Three things make up that judgment: intention, relationship, and listening.
Intention
A great voice performance isn’t just accurate delivery — it’s a series of tiny decisions about why a line is being said. Is this line reassuring the listener or challenging them? Is the pause before the word a beat of hesitation or a beat of confidence? That’s not pattern-matching. That’s a human being making a choice based on understanding the moment the way another human will experience it. It matters because those tiny choices are exactly what separates a read that sounds fine from one that actually persuades, reassures, or sells. That’s the whole job for an advertising agency or production company — a campaign doesn’t succeed because it sounded pleasant, it succeeds because it moved someone toward trust, action, or belief. Intention is what carries a voice from pleasant to persuasive.
Relationship
When I’m in a session, I’m not just reading copy — I’m building a relationship with the material, the brand, and often the director, in real time. A note like “give me more warmth, less salesy” isn’t a technical instruction. It’s a relational one. I have to understand what “salesy” means to this client, in this context, and adjust — not from a menu of preset tones, but from genuine interpretation. It matters because brand voice isn’t static — it shifts project to project, and someone has to be able to track that shift in real time. For an agency or production company, that relationship is what turns a one-off read into a working partnership — someone who understands the brand well enough that the notes get shorter and the results get better with every project, instead of starting from zero each time.
Listening
This is the one people underestimate most. Great voice work requires listening — to direction, to the emotional register of a scene, to what’s not being said in the script but needs to come through anyway. That skill is earned, not generated. It matters because the best direction is often the least literal. A director doesn’t hand you a setting — they hand you a feeling to chase: “Sound commanding, but not quite as commanding as a drill sergeant.” “Sound compassionate, but not sappy.” Those notes only make sense to someone who knows what a drill sergeant sounds like, what sappy sounds like, and exactly where the space between the two actually lives. That’s not a slider you can move. It’s a judgment call, made in real time, by someone who’s spent a career learning how to hear what’s actually being asked for.
AI can generate a voice. It can’t yet generate the judgment behind a great one.
What This Actually Means for Production Companies, Agencies, and eLearning Teams
This isn’t just a philosophical debate — it has real, practical implications depending on who you are and what you’re producing.
For production companies: The jobs where AI voice makes sense are usually the ones where nobody’s watching closely — internal placeholders, quick scratch tracks, low-stakes drafts. The jobs where a human voice actor earns their rate are the ones where the performance is the product: commercials, branded content, anything where audience trust and emotional accuracy directly affect the outcome. Knowing which bucket your project falls into before you cast (or don’t cast) saves you time and protects the final product — and it starts with specs that clearly communicate what the project actually needs, whether that’s an AI-appropriate placeholder or a performance that has to carry real weight.
For advertising agencies: Brand voice is relationship, not just tone. An AI read can match a style guide. It can’t build the kind of long-term vocal identity that makes an audience recognize and trust a brand over years of campaigns — the way a consistent human voice actor can become genuinely associated with a brand’s personality.
For eLearning developers: This is where I think the conversation is most nuanced. AI narration is tempting for high-volume course libraries, and honestly, it’s a reasonable call for some modules. But retention data consistently shows learners disengage faster from voices that feel flat or synthetic — and the modules where real understanding needs to land (compliance training with legal stakes, onboarding that shapes culture, anything emotionally sensitive) are exactly where a human voice’s judgment and warmth measurably improve outcomes.
The question isn’t “AI or human” — it’s “what’s actually riding on this performance?” Answer that first, and the casting decision gets a lot easier.
The Long Game — Where I See This Going
I don’t think the honest answer is “AI will never get here.” I think the honest answer is “not yet, and here’s why that gap matters right now.”
Sales taught me how people think. On-camera acting taught me how people feel. Voiceover taught me how people listen. Technology is teaching me how people learn — including how they learn to trust a voice, and how quickly they can tell when something’s missing, even if they can’t name what it is.
That’s the piece I don’t think generative AI closes anytime soon. Not the sound. The trust.
I’m not interested in pretending AI voice technology doesn’t exist, or hoping it goes away. I’m interested in understanding it well enough to know exactly where a real performance still wins — and making sure that’s where I show up.
So the next time someone asks whether a human voice can still compete with an AI-generated one, my answer isn’t going to be a defensive “of course.” It’s going to be the more useful one: it depends what you actually need that voice to do. If it just needs to say the words, AI has that covered. If it needs to be believed, that’s still a job for someone who’s spent a career learning how people listen.
I’d love to hear where you land on this — drop a comment and let me know whether you’ve tested AI voice tools yourself, and what you noticed. If posts like this are useful to you, follow the blog for more of these breakdowns as the industry keeps shifting. And if you’ve got a project on the table where the performance itself is the thing that needs to land — not just the words — that’s exactly the kind of work I love doing.