From it’s birth at Bell Labs in 1938, the Voder (an offspring of the Vocoder) sounded the alarm for voice actors far into the future. (check out a cool video that shows it in action) It wasn’t much of a threat then, but flash forward almost one hundred years and we voice actors are right to be concerned. The development of AI voice over is moving forward fast. The question on my mind, and most likely yours: Is AI voice over a threat or an opportunity? I’ll show you how it’s a little of both:
AI threat
The AI threat is real for voice actors who voice content for internal use by businesses. If you voice eLearning, Explainers, business presentations and any sort of on-line training, your lunch will be eaten first. Robots will come for this work first because it’s non-emotional for the most part. Scripts that are instructional or descriptive in nature and not meant to sell anything don’t require expression of a wide range of human emotion. They are the easiest for AI to emulate.
Once the voice over software is good enough for this sector, these jobs will move from humans to machines very quickly. It’s obvious why: The vendor can undercut human voice actors on price, plus there will be minimal charge if any for script changes or revisions. And Computer voices are always available, are never sick or complain and always sound the same, so there is instant matching of audio quality. It’s an audio engineers dream for “punch-ins”. The AI voice over software I’ve heard so far is pretty good, and all it takes for this kind of work is ‘good enough’. The AI threat is close to taking over this sector of work. If this is the majority of your voice over work, it’s time to diversify.
AI opportunity
AI opportunity is there for voice actors who can partner with developers as a source for voice over samples. You may think, Why can’t they just get Tanya from Accounting to say a few words and phrases into a mic then plug that into the AI voice over algorithm? Because Tanya from Accounting is not an Actor. To create believable sounding computer generated voice over, you have to sample human emotion in all its subtleties. That’s where we come in. But you have start making connections with developers now. AI opportunity will abound for voice actors who can be sources for samples early on. One such developer called Sonantic is focusing on video games. Check out this video to see how far the tech has advanced from the early days of the Voder.
How does it work? Voice actors record several random lines of dialogue, words or phrases and the AI goes to work emulating the voice model. The algorithm can then use that model to generate new combinations of words or phrases. Besides getting paid for the session AI opportunity for the voice actor comes with generating passive income every time their voice model is used. How much would a voice actor stand to make, and will it be supported by SagAftra? No word from the guild on that yet.
When you should worry
The general consensus on when you should worry is: not at the moment and maybe not for awhile. The one major weakness of present AI voiceover is its inability to express genuine human emotion. For that, you need a point of view. Humans can understand the purpose or meaning of a script. Computers can’t. That development is seemingly a long way off. Humans can hear real emotion a mile away; something a computer can’t deliver. Without emotion, we become disengaged very quickly
What to do now
You may be asking yourself – possibly through a veil of tears – what to do now. First, Don’t panic. Second, get creative. That’s one advantage you have over AI. For now. Diversify your client base in different sectors like Commercial, but mostly for celebrities and those on national SagAftra work. The lower echelon on Commercial work like regional car deal spots will be more vulnerable to AI voice over take-take over. Voice over work less vulnerable at first will be Audiobooks and maybe Video Games early on. Any type of voice over that leans heavy on acting chops and emotion will be tougher to crack. The human ear is incredibly good at discerning subtitles in human vocal communication. Two hundred thousand years of human evolution and communication present a formidable challenge to AI voice over.
So the answer to the question, is AI voice over an opportunity or a threat, seems to be: A little of both. Remember that old sarcastic chestnut from George Burns; “Sincerity. If you can fake that, you’ve got it made.”? AI voice over has said “challenge accepted!”. Now, about sarcasm. Can a computer express that? Uh…sure.