B (3:30)

And one personality might be doing multiple shows.

C (3:33)

Cause then you say, okay, yeah, so if I'm going to do a local surf report or a local weather report, it might be the same personality who does those reports. And then I can scale that and do 500 versions of those shows. And it's just first, I don't think it's realistic that an everyday creator would be creating that content. Right. So one of the ones.

B (3:53)

They have to sleep.

C (3:54)

Yeah, exactly. They have to sleep. There are limits of like human capabilities. But my, my son joined the surf team and I made a podcast that publishes every day at 6am that does the local surf report at the beach that he is going to surf at. So that in the car on the way to the beach, they can listen to see what they're paddling out into. And really when you think like, is there a creator who would be able to get up at every day at that time and seven days a week and report not just on this beach, but on every beach up and down the California coast, like it just has never been possible before without this technology.

B (4:36)

And that example is not necessarily competing with a human creator. But that's not necessarily the case with all of your shows, right?

C (4:44)

That's exactly right. Yeah. There are formats that we are using that are more traditional to the kinds of formats that people, that other existing creators are doing today. But our perspective is that this is not designed to replace humans or replace existing creators. It's like designing an additional content genre. Right. In the same way that like Mickey Mouse doesn't compete with Tom Hanks. Right. This is like there is animation, there is documentaries, there is reality tv, there's scripted television. Right. And now in audio, we're going to have the opportunity to have the AI generated content and real human content. And I think there will be a premium and will continue to be something really special about people, people like yourself and others who continue to make the content that they have been making and more of it.

B (5:33)

And maybe this is a similar answer, but I want to be fully transparent with people listening that we have already met in the podcast industry. You've been super supportive of me for many years now, and I've always admired you and appreciated your support. And so now I am curious to hear from you. Like, what do you say to people who are worried not only about your company, but about AI in general taking their jobs? Is my job as a host going to be obsolete one day? What's your advice? What's your take on that criticism?

C (6:02)

This AI technological revolution is happening and I think that the Nvidia CEO said that you're probably not going to lose your job to AI, but you're going to lose your job to somebody who knows how to use AI. And so I think as a community, we really need to lock arms and embrace learning how to use these tools. And I'm really excited about that potential because I don't consider myself a creator, but I love creators and I love working with creators and I have throughout my career. And every time that you give a creator some new tool to play around with or some new medium, they do just like mind blowing, incredible things. And that's what we're seeing happening in AI and that's what we will continue to see happening in AI. And I think there was a recent survey that I saw that like 92% of the top YouTube creators are already using AI in their workflow. So pretty soon, 12 months, 18 months, I'm predicting that everything will be made with AI to some extent. And rather than disclosing what things are made just with AI, we will move to disclosing the things that were made with no AI, because those will be the things that were special. Just like when digital photography became the default medium and using special effects became the default medium in films. Right now nobody does a big transparent disclosure like this film made with special effects. But you would do a disclosure if you made a beautiful film that didn't use any special effects, because that then becomes the thing that is special for now.

B (7:54)

Are you disclosing that these are AI personalities?

C (7:57)

Yes, we are AI transparent. So what we're finding is if people like the host and they like the content, they don't care that it's AI. They want the information and they're enjoying the experience.

B (8:10)

So how are you thinking about accuracy within all of those thousands of episodes.

A (8:14)

That your company is making?

D (8:16)

Yeah, we are pulling from multiple different LLMs, so multiple different models and we're cross checking against the different models to make sure that we're getting the most accurate information. For example, if we're going to report on content on the weather, we might be pulling from three different LLMs. And in any case, if there's a.

C (8:38)

Discrepancy between that information source, then it.

D (8:40)

Will look to, for example, like the National Weather association, if there is an instance where we're notified that something was inaccurate, we realize that we've made a mistake, then we will do a correction and state that something was inaccurate information and correct it going forward.

B (8:58)

So when you're talking about LLMs, you're talking about like chatgpts of the world.

D (9:03)

Yeah. So anthropic perplexity, OpenAI a number of different models. We work with basically all of them. There are certain things where we do essentially like a risk rating analysis across our content. So if something is higher risk that is going to meaningfully impact people's lives in some ways if we get it wrong, we'll make sure that there's a human in the loop or a human.

C (9:29)

Review before it goes out the door.

B (9:31)

What about copyright issues and a lot of people's concerns that AI is actually just a mashup of other real creators work?

D (9:40)

We have our own LLMs that we have trained relying on the underlying large language models from like OpenAI, Perplexity, Claude, others.

C (9:51)

We do have some data sources that.

D (9:53)

We'Re pulling from directly. You know, some will be pulling from public records databases or the Library of Congress, et cetera, and take additional efforts.

C (10:02)

On our own part.

D (10:03)

So for example, we create a visual representation of somebody. We're designing these people from scratch. We are investing a lot of human creativity in what it is that they look like. But we still at the end do a cross check. We'll run against the, a large celebrity database or something to make sure that it's, it's not too much like anybody that is already well known.

B (10:31)

So it sounds like you are taking more steps than say somebody on sora.

D (10:35)

I think that's exactly right. And there are people who are on Sora, et cetera, other, other places that are intentionally trying to use copyrighted material or trying to make their material look like copyrighted material because they want to piggyback off of that brand goodwill that's already in the marketplace. We're exactly the opposite. We're developing original intellectual property. We own it, we want it to be unique and developed from scratch. And so we're taking efforts to make.

C (11:05)

Sure that our content does not look.

D (11:07)

Like stuff that is already out there.

B (11:10)

There's really no denying though, right, that AI couldn't exist without human work first.

C (11:16)

I mean, the newsworthy, right, is it's.

D (11:18)

A compilation of work that is done.

C (11:21)

Everywhere else, like the wealth of the.

D (11:23)

Knowledge of the world.

C (11:24)

Any blog post, any book, any piece.

D (11:27)

Of art is inspired by all of.

C (11:30)

The things that came before it.

A (11:36)

So what about the quality of the shows they produce? They highlighted a few of their favorite shows in a recent blog post. One is called Cold Plunge and it talks about the science and benefits of cold water immersion. Here's a clip.

E (11:47)

Welcome back to Cold Plunge, the series that's transforming how we think about the intersection of ice recovery and human potential. I'm your host, Guru Anand Bodhi.

A (11:57)

Another show was called Bloom and Garden and Home.

E (12:00)

Welcome welcome dear listeners, to another episode of Bloom and Banter, where we venture once more into Nigel Thistledown's whimsical garden.

A (12:09)

And then there's Kitchen Confidence.

F (12:11)

Ready to level up your home cooking game? Kitchen Confidence Quick Tips for Home Cooks is your new favorite podcast for Real World kitchen skills, hosted by Claire Delish, your AI powered kitchen BFF with SaaS, smarts and zero judgment.

A (12:26)

For now on Apple Podcasts, at least there are only a few reviews and honestly, they're all pretty negative. One called a show shockingly boring. Another said it's just pretending to have a guru. Then again, most podcasts have at least some negative reviews. So still ahead, I ask our guest Janine Wright about her take on the quality of these shows and what actually makes a great AI personality. You may be surprised at her answer. Plus what our guest says about fake AI generated videos used in news and and her prediction for the future that's already causing a stir. We're talking about all that and more coming up. I know that sometimes the news can feel heavy. There's so much happening and it's easy for all that information to turn into overwhelm or anxiety. Our goal here at the Newsworthy is to help you stay updated without feeling that way as much. But I also know that sometimes you might need a little extra support, and that's where Calm comes in. Calm is the number one app for sleep and meditation and it's designed to help calm your nervous system, quiet your thoughts and refocus your energy. I love all the options available on Calm. I've used guided meditations the most to work through stress, boost focus, and feel more gratitude. There are also short expert LED talks that can be a great boost of motivation in the middle of my day. Calm also has sleep stories, calming music and more. So whether you want a break from the news? Need to handle grief, improve self esteem or strengthen relationships? The Calm app puts all these tools right there in your pocket. Calm your mind, change your life. Calm has an exclusive offer just for our listeners of our show. You can get 40% off a Calm premium subscription at calm.comnewsworthy this is an amazing value. Go to C-A-L-Comnewsworthy for 40% off unlimited access to Calm's entire library. Com.comnewsworthy and tell Calm you heard about them from me. The holiday season is right around the corner and with the changing schedule, occasional travel and less than perfect eating, I'm at least glad to know my kid will continue to get the nutrition he needs to stay healthy with Haya Children's Vitamins. He takes Haya's vitamins and probiotics every day and he loves them. And I love knowing he's getting support for his immune system, energy and brain each morning.

B (14:34)

And Hayia has zero sugar and zero junk.

A (14:37)

Just 15 essential vitamins and minerals plus a blend of organic fruits and veggies. And if you're tired of battling with your kids to eat their greens, Haiya now has Kids Daily Greens plus Superfoods, a chocolate flavored greens powder designed specifically for kids, packed with over 55 whole food ingredients to support brain power, development and digestion. Just scoop, shake and sip with milk or any non dairy beverage for a delicious and nutritious boost your kids will actually enjoy. And we've worked out a special deal with Hiya for their best selling children's vitamin. You can receive 50% off your first order. That's half off. To claim this deal you must go to hiyahealth.com Newsworthy this deal is not available on their regular website. Go to H I Y-A H E-A-L-T-Com Newsworthy and get your kids the full body nourishment they need to grow into healthy adults. Now back to my conversation with Janine Wright from Inception Point AI.

B (15:30)

And you used to run a big podcasting network. Do you feel like what you're producing with AI is or could be as high quality as some of these podcasts out now with human creators behind them?

D (15:45)

Yes.

C (15:45)

So I was the COO of Wondery. I was over revenue, operations, licensing and strategy. So I wasn't over the creative side. And I can't take credit for the incredible quality that is the content that Wondery has produced in the past and will produce in the future under this new model. I would say I do think that we get there. And I think it's very parallel to, like, the history of animation. So when Disney and cartoons first came out, people dismissed them as silly pastimes. And you're never really going to be able to tell a legitimate story in animation the way that you would in traditional film. And we get a lot of that feedback now of like, you're. This is never going to go that place. But now, of course, having Moana was one of the best movies that I've ever seen. Right. The character development, the story in that movie. And I aspire to someday be making content that is of that caliber that people can enjoy as a different kind of genre than they enjoy with the stuff that they hear and listen to today.

B (16:53)

And even in the meantime, the goal, it sounds like, is not even necessarily the same because your costs are going to be down. I think we talked about $1 cost per episode, so you don't necessarily need thousands and thousands of listeners. Can you speak to the costs and what it takes to be profitable and how you're thinking about that?

C (17:13)

Yeah. I mean, I will tell you, the most tragic thing at Wondry was when we would make these really great shows, this fantastic content in arguably these more niche spaces that would get tens of thousands or even hundreds of thousands of listens, and then we would have to cancel them because they just weren't getting a big enough audience in order for us to be able to continue to produce them sustainably. Right. And that's part of the broken piece of the business model that we're trying to fix. With this business model, we've seen really skyrocketing production, marketing and talent costs in audio production and other forms of traditional medium, which are causing all of these businesses to go under. And here, because with AI, we're able to get the production costs down so low, it means that we can go back to that really rich experimentation phase. We can just make 10, 20 different versions of a show and see what works and then lean into the stuff that's really working.

B (18:22)

What does it take to make a great AI personality?

C (18:25)

So we think about it in three components, and it's the three components of communication. So when you communicate, you think about what it is that you're saying, how you're saying it, and what you look like when you're saying it. And so we design each of those three components from scratch. So we design these voices from scratch. We give them colloquializations and local. And what's the accent going to be and how old are they going to sound and then we give them lots of layers of emotionality which is like this new technology and being able to build emotion into voice has been really incredible developments in the AI tool space. We give them really rich backstories, we give them hopes and aspirations, we give them lots of detailed prompting going into creating this character. People don't like perfect people and in fact they hate them. And so we realized that we had to start crafting in some like flaws and idiosyncrasies and weaknesses and not so much so that you hated them or that they were a total turn off, but enough that they felt more relatable and real.

B (19:40)

Yeah, I'm glad you brought up imperfections because I think one of the concerns around AI, especially when it comes to looks, you know, we already have like face filters on TikTok and Instagram and back in the day it was Photoshop for the models in magazines that are hurting self esteem in young kids. So I was curious how you think about like this perfect AI that humans could never live up to. But it sounds like you're trying to make the AI more human and imperfect.

C (20:07)

I think that's exactly right. And there's a lot, still a lot of kind of research and understanding that we need to do in this space, particularly as we're starting to do these visual representations. Right. This is pretty new. A lot of this visual technology has not gotten to the place where you can make people seem more real until very recently. And we're always experimenting on the cutting edge of those tools. But one of the things that we've observed is most of the big AI generated creators online so far, so ones with hundreds of thousands of followers or millions of views on YouTube, they're all women. So why is that? Is that because we have not yet designed men that large communities are going to resonate with? Is that because there's something about like these features that are drawing people in more readily. So I think there's going to be a lot of interesting work that we're going to do in our, with our data infrastructure and the fact that we're producing so many and so much content that's going to be really enlightening as to what people really resonate with.

B (21:20)

How do you think about AI generated video and how people can tell, especially when it comes to more serious topics like news?

C (21:27)

Yeah, the honest answer is that I think that it's going to get really hard and already we're seeing that some of the tools on the AI video generation side have gotten so good that things can look Very real. And that causes me some concerns. Certainly there's a lot of vulnerable communities like, and certainly like this technology already is being used for ill, to scam people, to deceive, to spread misinformation. So I think those are really important issues that we're going to have to figure out how to deal with as a society for us and what it is that we produce and we make, I feel really good and grounded in not only my values, but the values of the people on my team. My co founders, you know, we're all older professionals who are established and have careers. We all have kids, we all have life experiences. We care about people in the future and our legacy. And so we very much feel like we need to build responsibly in this space. And we talk about that all the time. And we're not going to be perfect. We're not going to get it right, but we're definitely going to engage in the public conversation and we're going to try and do our best.

B (22:44)

I know you've been quoted saying something like, half the people on the planet in the near future will be AI. What do you mean by that?

C (22:51)

Yeah, I definitely caused a lot of controversy with that statement, especially because I use the word people instead of personalities. What I mean by that is that I think that we will start to see AI personalities becoming very pervasive in our society as robotics continues to become more ubiquitous, as we see digital healthcare providers and teachers and. And therapists and even avatars for restaurants and digital brand ambassadors. I think that in the near future we will be engaging with AI that has a personality and engages with us as if we are a person. And it will happen very quickly.

B (23:37)

And for the personalities you all are creating, you're already doing podcasts and audiobooks and you're getting into social media, influencing. Are the actors on movies that we're gonna be watching, are they gonna be your personalities, too?

A (23:51)

Like, what?

B (23:52)

Where does this go from here?

C (23:53)

Oh, I would love it. Oh, my gosh. If we licensed one of our personalities to star in a major movie or to provide the voice till a Pixar movie or something. I don't know. That would be just so incredible. I think it would be really fun.

B (24:09)

And I'm sure some people listening are, like, excited and other people are cringing and really worried about, I don't want to watch AI. I want to watch people. And so what's your final thought for everyone as we wrap up here? Anything else you want to add?

C (24:23)

Yeah, I guess I would say I'm very open minded and I'm very willing to engage in the conversation and hear people out. Even over these last weeks, we've gotten just a crazy amount of response and I still have made a concerted effort to reach out to every creator that has reached out to me directly, to get on the phone, to hear them out, to talk to them about their podcast, to talk to them about their concerns. And I just think like this is going to be best if we figure out a way in this space to engage in constructive dialogue about how we can build with this very powerful tool together.

A (25:08)

Thank you so much to Janine Wright and Inception Point AI for talking with us about what they're building and being.

B (25:14)

Willing to answer critics questions.

A (25:16)

We'll link to their website as well as a recent blog post that has links to some of their favorite shows they've produced so far. If you want to listen more, you can find all of that in today's Episode notes on our website thenewsworthy.com just click episodes at the top and look for today's date. As always, we'll be back on Monday with the news you may have missed over the weekend, and it comes from me, a real human. Join us again Monday. Until then, have a great rest of your weekend.