The Hidden Costs of Using AI Voices

There's a loooootttt of talk about AI voiceovers, lately. Despite the fact that they take work away from us hard-working, lovable voice actors (cue sad puppy dog eyes), some brands have already made the switch from real voice actors to AI voices, thinking they would save money and time while increasing their sales.

But, the reality is proving to be a bigger headache (and much less cost-effective and profitable) than expected. There are a lot of hidden costs that most brands aren't aware of until it's too late.

Expert analysis shows that it typically takes several hours to manipulate, test and edit an AI voice to hit certain inflections and timing restrictions required by the script, not to mention tone shifts and pronunciations throughout. And what about syncing it with already existing video? Just thinking about the tediousness involved is exhausting.

What typically takes only 30-60 minutes with a real voice actor in a directed session (or a self-directed session), can now take hours if not days to achieve with AI. And, even then, it’s still not quite right. 

Then, of course, there’s the issue of whether or not you’re getting your AI voice from an ethical company. Some do compensate actors appropriately to use their voices for AI, but most don’t. And some even steal and resell actors’ voices. (And these companies don’t exactly promote the fact that they’re majorly under compensating actors or selling stolen goods.) So, it’s crucial to do the extra legwork (and take more time) to research AI companies to make sure their synthesized voices are built on fair, compensated partnerships—to avoid legal and creative risks, later on.

Creatively, AI has no ability to create its own point of view while reading a script. It has no understanding or interpretation of the words. That, inevitably, falls on the shoulders of the technician typing in all the commands for the AI voice.

So, while the software may be cheap(ish), the hours of manual editing to fix awkward pacing or pronunciation in a 30-second spot can effectively raise your overall production cost. Half the time, brands end up ditching the AI voice altogether and hiring a real voice actor anyway. So, as you can see, going the AI route doesn’t necessarily make things easier. Or, cheaper.

And here’s the kicker. Around 77% of consumers still trust human voices more than AI voices. (How many of us have clicked off a video the second we realized the narrator was AI? There’s just something icky about it. It feels almost violating, in some way.) 

Even though most of us can only detect an AI voice 51% of the time, according to Neuroscience News, the back of our brains instantly recognize AI. And that sends our brains on a mission to figure this AI voice out. Which means, consumers are no longer listening to your brand’s message; they’re too busy trying to figure out why the person speaking isn’t hitting their internal relatability center. In other words, they sense something is off. And once that happens, consumers decide that the brand producing this ad or narration must also be, by default, off.

Human voices activate the regions linked to memory and empathy, while AI triggers the parts of the brain responsible for error detection. In other words, if you use a robot, your audience is too busy "fact-checking" the audio to actually bond with your brand’s message.

Now, if your script is dry and robotic, AI will do a pretty good job. But, who wants to lead with a bland, mediocre script? If the AI voice sounds awkward and struggles to capture the script’s overall personality, then you most likely have a good script and should prrrrrobably get yourself a real human voice actor to record it, if you want it to have any real impact.

And the stakes are higher than you might think. Research from Adobe Express shows that 58% of consumers trust a brand less if they use AI voices in their ads. And other research shows that sales increase the more personality and overall humanness an ad has. 

The second people catch on that it’s an AI voice, it actually lowers their trust of your brand. It cheapens the experience. It leaves them wondering, if the brand’s voice is false, are their promises false, too?

On the flip side, 51% of consumers are perfectly comfortable with AI voices in video games and 45% are cool with it in customer service recordings. However, when it comes to things like news, podcasts and advertising, that’s where consumers draw the line, with only 9% being ok with AI voices in news and 8% ok with it in podcasts. That’s pretty bleak.

People don’t want to give up their hard-earned money for a company they don’t feel they can trust. And trust is everything in this business. Hire a human, and you’ll get a read that moves people (to buy your brand). Write for a robot, you get a bunch of viewers scrambling to hit that Skip button.

When the script is good, hiring a human voice to add that relatable sparkle will be ten times easier, faster, and frankly, more profitable in the end. (Not to mention way more fun.)

AI certainly has its place in our industry, don’t get me wrong, and I’m all for ethical use of it, but the results are pointing to a downturn in sales and consumer loyalism for those who use it in their ads, podcasts and journalistic videos. So, why risk it?