Jim Freeze Hi and welcome! I’m Jim Freeze and this is The ConversAItion, a podcast airing viewpoints on the impact of artificial intelligence on business and society.
[UPBEAT MUSIC]
The ConversAItion is presented by Interactions, a conversational AI company that builds intelligent virtual assistants capable of human-level communication and understanding.
In this episode, we’ll discuss AI and voice as it pertains to the entertainment industry. Answering questions like: What does the emergence of voice technology mean for voiceovers and voice acting? And: What can technology companies learn from the entertainment world when casting an AI-powered voice for their brand?
We’re joined by Jim Kennelly, the owner of Lotas Productions, an audio recording studio in New York’s Flatiron District. Jim has over 30 years of experience finding the right voice for all kinds of media.
Jim, welcome to The ConversAItion!
Jim Kennelly Thank you, Jim. I’m very happy to be here. Great to join you!
Jim Freeze Yeah I was actually really looking forward to this one for many reasons. We’re going to dive into some really interesting topics. So maybe we could kick it off—you could detail how you’ve seen, over the course of the past few years, technology impacting the voiceover world?
Jim Kennelly Sure, it’s a great question. One of the biggest effects of technology in the voiceover industry right now is speed – the need to update programs, to update commercials, to share information with your clients, to share brand information. You get data back so quickly that they’re able to analyze the effect of a commercial, an e-lesson, some sort of a voice prompt, they get that information back so quickly and analyze it that they can update it. So speed is the number one thing that is affecting the voiceover industry—speed and point-to-point technology are the biggest game-changers right now.
Jim Freeze Interesting. So how have you seen AI, in particular, impact the industry?
Jim Kennelly Well we always make the comparison between AI coming into society the way radio came into society in the late 20s, early 30s. You have this technology—radio—that everybody knew worked, everybody was very happy to put it into their homes. But then the problem became, “What do we do with it?” So there was a generation of men and women who built the foundation of this industry. They were given the responsibility to come up with programs, to come up with shows. That’s the big influence that AI is having now—it’s giving a whole rebirth to the voiceover business because obviously, for actors, and creative people, for writers, directors, producers, we see them coming into the industry because they’re going to have to write conversational dialogue and that’s going to require some different talents.
Jim Freeze So could I assume from what you’re saying that you think the advent of AI is going to have as big an impact as the introduction of radio did?
Jim Kennelly Oh absol—without a doubt. Particularly what’s unique about AI, where it’s really heading down the road—there’s going to be so many individual personalities that are going to be needed to communicate. Obviously radio stations have disc jockeys and certain personalities, but brands are going to need to know that they’re culturally relevant in every market. Consumers today are so used to specific, specific responses—a variety of specific responses.
Jim Freeze I’d really be interested in hearing a little bit about the process of casting voices. What do you consider when casting voice actors for projects in the entertainment world?
Jim Kennelly Casting is—starts with a conversation. We have to talk to the brands, see who their audience is, who they may already know that or maybe they’re trying to find an audience. So we have that conversation, we start to understand, we get words from them like we want it to be conversational, or we want it to be millenial-like or we want it to be like on a podcast. Those words lead us to the type of actors we’re going to go to.
Jim Freeze It sounds like there are certain qualities of human voice that are particularly impactful in the context of brand representations or whatever message is trying to be delivered. You know there are so many distinctive voices. I think about someone like James Earl Jones as Darth Vader and how that voice is – it’s more the persona than than kind of what the character looked like. So I’m wondering, are there things like tone, depth, accent, cadence of speech, that you kind of consider when you’re casting? And how do you know which attributes are right?
Jim Kennelly We do factor exactly those things in. How quickly do you want the copy to be read? Or how slowly do you want the copy to be read? Is it an anthem piece and is it going to be played in a really large room one time? Then you may want a James Earl Jones type voice that slows the copy down and brings the importance to this message. Or if it’s Raymour and Flanigan, which is a big furniture store in the Northeast, they may want something quick and rapid that says delivery in three days.
Jim Freeze I have to say, listening to you talk about that – when we build virtual assistants for for customers of ours, we kind of work with them on personas and we think that the voice, the language, whether it’s a woman or a man, we think it very much should reflect the kind of brand attributes they’re trying to reflect. And so we do think it’s important, and customers take it very seriously and we really advise them to think about the persona they’re trying to manifest.
Jim Kennelly Oh I think that’s absolutely right. As consumers, we’re groomed to expect like boundless options. So it stands to reason that the range of AI personalities in the future is really going to explode. But one of the things that’s holding AI back from that right now is that it requires a lot of intensive manual effort to create these personas, these sonic identities, to make them have a conversational feel. What you’re trying to do is make the personality be approachable, so the whole trick is to get this–these machine-learning AI systems to interact with people because then you get more data. The more that people interact with it, you get more data. And the system can’t learn if it doesn’t get to interact with the human being that it’s supposed to be working with. It just takes time.
Jim Freeze So you just touched on a topic that I wanted to talk about because I read an article you wrote in Backstage Magazine that relates to sonic identities. And you wrote about the history of voice in entertainment and the recent impact of AI on its trajectory. And in it, you mention that as people become more comfortable with AI assistants, brands will be looking for voice actors to build these notions of the “sonic identity”—the “voice” of a particular brand. Like Siri for Apple, or Alexa for Amazon.
Snce you kind of wrote that article, have you seen the concept of sonic identities impacting voiceover and voiceover acting?
Jim Kennelly I think we’re just starting to see it. When we talk to companies that want to create these personas, they’re not like advertising agencies or production companies of the past who had people who were very familiar with casting and had people on staff who would bring talent to them. These are companies—I’ve been at a number of conventions and you sit at a table and they’re like, if we give you copy can you get us some thirty year old women who sound like mothers? Can you do that for us? And I’m like I can do that for you before I even get out of this seat. I could do that within the hour, I can have voices for you. So there’s a whole world of production companies that are a little more on the engineering side and they’re maybe not as involved on the artistic side. So the voiceover community, we’re going to bring our artistic expertise to these engineers and then they’re going to start to create these conversational user interfaces, these CUIs, that people are just really going to dig. They’re going to love this experience.
Jim Freeze Do you think there is anybody out there yet, who is doing a good job of starting to develop a sonic identity out there? Is there a brand out there yet that you kind of think of as a gold standard?
Jim Kennelly I don’t know if there is a gold standard one. You see a lot of flash briefing or skills on Alexa—and have jokes and there’s things like that. But I think what we’re seeing right now is really like the top layer, it’s the beginning of it. And as people become more like conversation, like what we’re able to do right here where you and I have cues and we understand how to have a conversation—we stop and start. They’re going to bring voice actors in to create that experience, but then slowly the engineering underneath is going to fade away. And as it becomes invisible, it’ll really feel like a conversation. There’s a lot of response-type programs out there but we’re really interested in the world that goes deeper than that.
Jim Freeze So that’s interesting. We see things similar in our business. Our conversational AI virtual assistant replaces something called speech-enabled IVR. XX
It’s not uncommon when we roll out an initial application and you go from telling people what they’re allowed to say, like “You can press 1 for this or 2 for that” and instead, it’s an open-ended questions like “How may I help you?” What we often find is that the intents are different than what they thought when they initially developed it. Does that surprise you?
Jim Kennelly No it doesn’t surprise me. The responses that people give to these smart devices, they’re always challenging them. They like the fact that they have personalities so I think it’s very common that like you think your client’s going to go right with your brand and all of a sudden they go left with it. That will happen a lot, but again that building and testing is what’s going to make the voiceover business explode again as we look at three, five, ten years down the road.
Jim Freeze Yeah and you hit on another key point which is it’s about data. AI is all about data. It’s an iterative – at least in our experience – it’s an iterative process and you learn as you roll something out and it gets to the point you made earlier about speed. Speed is so critically important to be able to adapt quickly because when you have data, you can learn pretty quickly.
Jim Kennelly Right and if you don’t keep people entertained, they’ll just drop, they’ll drop the system. That’s why you need voice actors to deliver the nuance that will hold people’s attention like we touched on. They’ll start to use it more and you’ll gather more data. And as you gather more data, you can even program back the responses even better. So that’s why you have to—that’s why the voice actors will be an important part of it.
You know there’s a place for a robotic voice that’s straight up if you’re saying “The next flight to Philadelphia is at Gate 629.” Well that’s nice, I don’t need a dramatic voice to tell me about that. But if you’re going to start to tell me about what types of music that I might prefer, and now you know I like jazz and we go a little deeper into jazz or the type of movies I might start to enjoy, here are some suggestions, you’re going to want your friend to say that.
Human conversation, you don’t like boring people. So obviously you don’t want your AI to be boring when you deal with it. You want it to have an exciting personality sort of to match yours. The ultimate goal is to make the personality of the AI to be like the consumer who’s using it.
Jim Freeze Yeah, so, I was thinking of asking you a fun question and so here it is: I’m wondering if you think I have a future as a voiceover actor?
Jim Kennelly Absolutely. Absolutely. Everyone does because we’ve entered the world of voiceovers now where, when I started it was a world of like deep, strong male announcers and when EF Hutton talks, people listen. And that was important and there were variations on that, but basically that was the voiceover world. But as we came into the 2000s, we see a transition out of that.
We’ve had a whole world of–you know the last ten years its been all about conversational reads, conversational reads, which again was pointing towards AI. But now we’ve actually moved beyond conversational reads and now we’re starting to see where they’re looking for millenial-type reads, they want a podcast-type of read.
So we’ve spent some time recently looking at people who do do podcasts, not necessarily our voiceover actors, but they do podcasts because they handle large amounts of information and they talk in this very relaxed, cool kind of chill way. And this is a voice that brands are asking us for. So no voice is ever ruled out. Your voice is in Jim.
Jim Freeze Alright, I’m glad that I still have potential in that career. So, one last question for you: In a lot of industries, AI is creating a fear that it’s going to displace jobs and takeover jobs that humans are doing today. Do you see that apprehension in the entertainment industry at all?
Jim Kennelly Yes, you do see that a lot. There’s a fear of synthesized voices. Obviously the AI is going to slowly be able to respond if you have this wealth of captured sound—words and sentences—but I think they’re still going to need voice actors to create the nuance, to really develop a character, to even help create the illusion of a conversation. That’s always going to be needed. It’ll still be a response–you know speak-and-response-type experience but I don’t see it taking jobs away from the industry like in–far, far in the future you might be worried about that but certainly not in the next ten years. They’re not even close to that kind of technology.
Jim Freeze Well who can even predict beyond ten years what’s going to happen? I can’t predict what’s going to happen next week. Hey Jim, this has been fantastic.
Jim Kennelly Thank you, it’s been my pleasure also. You know, we’re really looking forward to a world where we’re we really just talk to our smart speakers, and not at them. I think that’s what is coming next for the voiceover industry and it’s going to be fun to be a part of creating this brand new industry just like men and women in the radio years in the 1930s did it. Now it’s our responsibility to do it and we should have fun being creative, bringing art and science together. That’s a good occupation.
Jim Freeze It is. I really love the analogy of the adventive of AI being as impactful, or potentially even more impactful, than the introduction of the radio. That’s a great thing for us to think about. And I want to thank you for your time. This is fantastic and best of luck!
Jim Kennelly Okay thank you Jim. Best of luck to you. And enjoy the rest of this–the rest of the podcast.
Jim Freeze Thank you very much. Bye, bye
[MUSIC PLAYS QUIETLY]
On the next episode of The ConversAItion, join us for a discussion on AI, automation and the future of work, with prolific author and Boston-based Professor Tom Davenport.
This episode of The ConversAItion was recorded at the PRX Podcast Garage in Allston, Massachusetts, and produced by Interactions, a Massachusetts-based conversational AI company.
That brings us to the end of this episode of the ConversAItion. I’m Jim Freeze. Thanks for listening, and we’ll see you next time.
[UPBEAT MUSIC]
