A new startup says they are about to take A.I.-generated voices up a notch by giving the tool the ability to feel much more real.  Zeena Qureshi, CEO of Sonantic, joined Cheddar News to discuss further.

None of the Above 

New Method to Make A.I. Generated Voices More Expressive

Welcome back to the show in the past few years, artificial intelligence has become an integral part of our daily lives. It's not just a voice assistant like Alexa or Siri, the technology has permeated nearly any daily task you can think of but what if Alexa could do more than just turn on your lights or play some music. When ai voice startup says it's about to take things up a notch by giving generated voices, the ability to feel much more real, so real. They can express subtleties like teasing even flirtation. An article in the verge a few days ago caught a lot of people's attention. It was called listen to an Ai voice actor. Try and flirt with you. You've got a lot of people a lot more familiar with the company we're talking with today. Take a listen. Hi there. I'm not really sure how to start this. I wanted to talk to you about something something we're all looking for. I want to talk about love. I get it. It's kind of awkward. How do you connect with another person finding that common wavelength and really connecting? Mhm. A connection that could turn into love? Xena Karachi joins the show Now Xena is Ceo of semantic. Welcome to the show. Nice to have here. So why was the emotion of flirting specifically something that you and your colleagues at least wanted to try to recreate. It certainly creates a lot of buzz. A lot of talk that was a very heavily trafficked piece for the verge a few days ago. Thanks JD It's lovely to meet you. Yeah, for us we um we actually designed this for a valentine's day lunch and we got some inspiration from the movie her. Um and what we really what we really do is we create human realistic voice um for gaming and entertainment. And so we've created this Photoshop for voice. Um and the art is really in the subtleties of performance and so flirting falling in love, being able to convey. That is actually really difficult. It's even hard for human beings to do that. Um so that's why we went after that. So help someone who's watching this better understanding may not have a ton of familiarity with Ai who says, okay, well sure. I mean it's one thing if I take out my iphone and ask Siri to open up a task or a calendar but I have no frame of reference by which a more emotional Ai would be necessary. What do you envision this being necessarily as to why it's so important. So for us we work with entertainment studios primarily triple a gaming studios, film and tv and for them immersion is really important. You need to be able to hear that voice and really understand it as it goes through the story. And so for us we really look at all those different nuances. Um if the story doesn't sound good, you can't really relate, you don't really get immersed in that. And so what we do is we built this Photoshop for voice to allow audio directors and narrative designers really edit sculpt and director performance on demand. This helps both actors and studios studios in a way that they can create performances really quickly. Iterate quickly and actors um for them to be able to Work on multiple projects at once. Um they get a profit share every time their voices used, their voices preserved the AI can shop for them 24/7 or flirt for them. That's amazing. Now, Symantec also worked alongside a very well known actor. Val Kilmer of course, helped him reclaim his voice after throat cancer. Two story, a lot of people know but may not know the semantic side of it. I believe we have a clip, we're going to play that and then I'll ask you a question about it, Xena on the other side. But now I can express myself again, I can bring these dreams to you and show you this part of myself once more. A part that was never truly gone. Just hiding away Xena. What were the unique challenges or opportunities about taking on a unique case like that? Well, this was actually being able to give al his voice back was an incredible project. And for me um working in tech is great, but it's it's really nice to be able to give someone um that type of impact and it's really special because Val is such an incredible performer himself and so really looking at all those nuances and being able to give him that ability to interact on a day to day basis with his, you know, his original voice. And then being able to let him work with that voice if he wants to later down the line. Um for us it's really about creating that impact and being able to give that back to him. Yeah, there's something so incredibly beautiful about that, especially for people that that have followed the career and the challenges there Val Kilmer in the last decade or so. Um so far what are other types of customers here that that you've been working with as well outside of some of the more high profile ones like we just played predominantly triple A video games. Um For them they have really long development cycles. So anywhere from 3 to 10 years and the average game is 30 to 60 times longer than the average film. So for them with these iteration cycles, they need a lot of volume. So we've worked with Obsidian entertainment remedy games just to name a couple. Um but a lot of our studios are under N. D. A. So we can't really announce them until their entertainment products come out. Can you talk to us a little bit more about any type of forward or future thinking ambitions with regards to where you could go if you could name explicit clients totally understand. But what are some, you know, even sort of North star goals. I wonder here of the company things that that maybe the rest of us couldn't even yet imagine for us we really want to be that C. G. I of audio. So if you look at C. G I, um synthetic media for visuals has been around for a long time and now it's the industry standard for all films, games, um, simulations. And right now we're at an inflection point with technology where you can do the same thing for audio. And so for us, we're really paving that way for artificial voices and making it sound as realistic and help as believable to really help with those immersive experiences. Xena Karachi is the ceo of semantic. Thank you so much for taking the time. Congratulations on some truly awe inspiring and incredible and as we just saw their very important work in the lives of many people, Xena thank you for being here.