Generative AI in Modern Game Design with Elliot Wolf

Welcome back to On Production, a podcast brought to you by Wrapbook. Today, I'm exploring the cutting edge of storytelling where AI meets true crime and narrative gaming. Our guest today is Elliot Wolf, who is the executive producer behind Prime Video's On Call and the son of legendary Dick Wolf. But as co-founder and chief creative officer of Wolf Games, he's launching a new game, Public Eye, this summer. It's an AI-powered daily murder mystery game that lets players team up with an AI detective partner to unravel new cases each day. Super interesting. I wanna know though, what does it mean to reimagine crime storytelling through generative AI? What are the possibilities? How does this innovation maybe translate to TV podcasts and beyond? Because Elliot knows that world as well. And with that, welcome Elliot.

‍

Elliot Wolf (00:49.859)

Thanks for having me. Excited to be here.

‍

Cameron (00:52.034)

So yeah, man, this is super interesting. A lot of the guests that we've had on the show, across the production spectrum are all bringing AI into their workflows or thinking about how AI is going to impact their workflows. And so it's not normal for us to have somebody in the gaming community, but obviously your background really spans both. And it's interesting to see how AI is impacting all of media, heck, even at Wrapbook.

‍

You know, we're using AI and these workflows to make them smarter, more intelligent, more interesting. But I'm curious, you know, from TV to gaming, what inspired you from your TV-centric background to launch an AI-driven gaming studio?

‍

Elliot Wolf (01:33.997)

Yeah, I've always been a storyteller first and foremost, regardless of the medium. And I've long felt that there's a massive opportunity at the intersection of television and gaming. And if you look at interactive television to date, it's been extremely hard to produce because you have to create four to five times as much content for a single episode. And that's only to allow for the most basic choice, right?

‍

six, 10, 12 choices throughout an hour long episode. And it's somewhere between a lean in and lean back experience where you don't fully get to relax like you do and unwind watching your favorite TV show, but you're not engaged enough, like a traditional game to the point that I think the experience fell short for most people. And what really excited me looking around the corner with

‍

what I was seeing with AI and technology was the ability to produce that content in a manner that allowed for infinite choice and enough variability so that the viewer or player was really engaged second to second in forming their own path through the story. then as I started to really play with the technology, this

‍

bigger opportunity in my mind emerged, which was personalization to the player, where you can increasingly become tailored to that player's interests, that player's location, and so forth. Using Law & Order as an example, it's set in New York, but there's a ton of fans of the show in Chicago, Los Angeles, and Des Moines, Iowa. And traditionally,

‍

it's a broadcast show, you've never been able to experience Law and Order in the city that you actually live in with public eye and the experiences that we're building. You can play the story where you are, and it feels that much more personal to you because of it. So that's the first of, I think, a lot of personalization tactics that you're going to see emerging with AI production.

‍

Cameron (04:00.792)

So what was the process of actually getting this game built? What's the gameplay feel like day to day? And it's super interesting. So right, you're saying you have all of these new capabilities with these AI models. In the TV system, you've got all of this footage, but your story is quite linear. You're sort of blending classic storytelling with these generative AI models in some sort of user experience. What's it feel like? How does it work?

‍

Elliot Wolf (04:29.849)

So Public Eye, our first game, is a daily murder mystery game. And looking at the mobile market for crime games, there's quite a few really successful games out there. But what we found was the biggest pain point by far and away was the cadence of stories being released within those games. You can solve a murder mystery and then wait six to 12 months for the next case to come out with these mobile games that we've seen.

‍

And we felt like there was a massive opportunity for players to get into a daily habit of solving cases. And once you do that, you can create retention to the point that it becomes what I see as the most fun version of a daily brain exercise, where you're testing your deduction skills, your logic skills in order to solve a case.

‍

So the gameplay for Public Eye is essentially a deduction puzzle where you are moving through the facts of the case, the evidence of the case in order to deduce who did it, largely by eliminating who couldn't have done it by either matching pieces of evidence to find an actual clue or through interviewing suspects.

‍

It feels pretty immersive as if you're the detective moving through the traditional steps of solving a homicide.

‍

Cameron (06:08.59)

That's pretty cool. How long have you been in development on this game? Who'd you work with to make it happen?

‍

Elliot Wolf (06:15.641)

So we have an internal team of game designers who have been fantastic in building this game for a while now. Wolf Games is just over two years old. And we started by building the technology to support these games. And we've been developing Public Eye for a good amount of time now with the idea that the underlying technology of Public Eye will

‍

also allow us to build a lot more games across genre that have equal personalization and become increasingly tailored to them.

‍

Cameron (06:52.462)

I was looking at some press, Elliot, about the game and I saw that you were saying that it's your view that you don't think AI should necessarily make traditional content cheaper, rather allow you to create experiences that just simply were not possible before. What new creative potentials have emerged for writers, designers, game developers in this medium? What's the relationship between

‍

your writers thinking of cases versus the game player. Like what's the synergy between these two worldviews?

‍

Elliot Wolf (07:27.939)

Yeah. Well, firstly, I don't think there's ever been a more exciting time to be a creator in the entertainment industry. As someone who's created a TV show, the thing that I hear by far and away the most has been no, because ideas have to appeal to millions of people. And then studios have to decide that because of that, they're willing to invest tens of millions of dollars to bring your idea to life.

‍

And now a creator can meaningfully communicate their idea in a visual medium by themselves. And with that, you now need to find an audience. And for me, the, I'm a huge believer in a thousand true fans. If you can find a small audience that loves what you're creating, there's a business there. So,

‍

With AI, I feel like there is an opportunity to go after very specific groups of people, whether that be people with specific interests, from basketball to bird watching, whatever it may be, to people in specific locations. We can tell stories in New York for those who live in New York and create a much more immersive experience for those folks by allowing

‍

interactivity. Going back to what I was saying at the top of the conversation, the amount of interactivity that we have in our narrative games simply wasn't possible without AI.

‍

Cameron (09:10.37)

Why is that?

‍

Elliot Wolf (09:12.611)

cost. was unbelievably expensive to produce the amount of choice necessary to make it a truly interactive experience. When you introduce simplistic choice, you're already doubling, tripling your budget because you've got to go account for all that choice with additional content. And for us, we are now able to allow for enough choice where it really feels like game play. And

‍

For our writers, I don't want to speak for them, but it's been extremely liberating because the idea of a branching narrative can actually lead you to different destinations in a story and have different outcomes.

‍

Cameron (09:58.894)

The prompting of how you write this is so interesting, I'm guessing, because it's like you take these different approaches as a player, but the underlying prompt still has to remember the context of who did it, right? Or, I mean, can you go totally off base? What degree of freedom do you have as a user, or can you just fail in solving the case on any given day?

‍

Elliot Wolf (10:22.541)

Yeah, so we are still on rails. So there is still a limited amount of choice that a player can make today. Obviously, we want there to be more more agency on the player side as we progress. But our technology, we build our own tech platform. And the underlying logic is key to our products. It's what we call our story core that's able to maintain narrative consistency.

‍

and maintain character consistency and most importantly, a singular timeline of events. So that underlying logic is there.

‍

More broadly, we're still not prompting necessarily a story. We're having writers come up with an idea. And that idea can be something that's extremely relevant, right? We can take something that's trending on the internet. Let's say the Met Gala is tomorrow or the NBA Finals was two weeks ago, whatever it may be. We can create a story around those events. Let's use the NBA Finals as an example. After game seven,

‍

We did a beta case around the NBA finals that incorporated information about the winner and so forth. And we are able to stay in the zeitgeist or in the internet conversation. That decision is still coming from a writer's point of view, right? Someone is sitting down saying, this is a story I want to tell, and then leveraging our technology and working with our platform to bring that story to life very quickly and hopefully soon almost instantaneously.

‍

Cameron (12:05.474)

This is really interesting Elliot, like, cause again, you have this TV background, you come from a TV family, you've seen how the system works, what resonates with audiences, what doesn't, but you're also totally right. We're also living through this incredible tsunami of AI, seeing how it is so creative and destructive and good and bad and all these different things. When you look ahead three to five years, how do you imagine AI powered storytelling evolving across all of entertainment, not just games, like around personalization, interactivity?

‍

new business models. mean, you talk about a thousand true fans. It's like a brilliant idea. You know, you only need a thousand true fans. You don't need to have massive millions of member audiences anymore, which is a totally valid point. You know, I've even thought, you know, like reality television is radically popular. They have to get rid of most of the footage in order to have a story. But like, what if you could interact with your very favorite characters?

‍

in all these different paths. mean, what do you, how do you think that what you're building in gaming is going to transcend or interact with, with, kind of classic, film based storytelling.

‍

Elliot Wolf (13:14.755)

Yeah, well, first off, I am a huge fan of TV, both scripted and reality. I don't think that goes away. I do think that there is an increased amount of choice in how fans choose to interact with their favorite IP and their favorite IP universe. For us, we see our products today as a means of touching your favorite IP every single day. Now,

‍

There is a ton of value in a television series in that scenario still because you still need to build affinity for that IP. Right? So let's use any given streaming show as an example. You're doing a eight to 10 episode run largely with at least a year and a half in between seasons, but you're building a huge fan base with those 10 episodes.

‍

That fan base is left waiting for the next season after they finish the 10th episode. But what if there was a world where you can interact with that IP universe every single day with new stories in the universe, either touching or not touching the main characters from that show? That's what we're really interested in. And I also believe

‍

Cameron (14:41.472)

It's kind of a, yeah, it's a really cool idea. Like even you've seen like the squid game, how it pushes into the reality competition back into a new season. You could imagine even in a game playing that.

‍

Elliot Wolf (14:52.153)

It's an IP ecosystem and the temple offering will continue to be the TV shows. But the means of interacting with that IP on a daily basis can come from smaller snack bites of content, interactive content that have game mechanics that allow you to build status within that IP universe, socialize your success.

‍

within that IP universe. And for the streamers, it's really enticing too, because they want to get people coming back every day. They want people to stay on platform longer. And this is an amazing way of doing so because you're working with the IP that they know that their audience already knows and loves. So we're really excited about working with...

‍

existing IP and extending those IP universes alongside building out originals.

‍

Cameron (15:54.808)

That's really cool. I'm curious, Elliot, what's the most memorable user AI moment you've seen in testing?

‍

Elliot Wolf (16:02.734)

user AI moment, as in how a player has interacted with the AI.

‍

Cameron (16:07.702)

Yeah, guess it's an amazing thing to imagine. We've sort of passed the Turing test. Of course, you can set these guardrails with the characters that you're interacting with in the game. And they'll speak to you in these really convincing and elegant ways. But have you seen an interesting experience or sort of an aha moment between a user interacting with the game while testing it?

‍

Elliot Wolf (16:33.507)

Well, the aha moments largely come from the gameplay, right? It's actually the player having the aha moment with themselves saying that they discovered a clue or they solved the case correctly. I think what's been most impressive with the AI is how with proper writing, human and emotional, these characters can feel.

‍

That is still a team effort, right? We're refining everything that comes out of our system and fine tuning to make sure it's the best possible experience for the player.

‍

You've got to keep in mind the timeline of how these events have unfolded. Ten months ago, people didn't think video was going to be a reality for at least the next six, seven months. And look at where we are with video today. It's all moving so quickly that these moments that make people say, wow, are going to be increasingly common, I think.

‍

that we're already getting to the point where...

‍

The experiences that have been unlocked by AI are, in my opinion, really eye-opening and fun. But tomorrow, what I'm really excited about is that ecosystem again, where you are watching a TV show that has traditional production, moving over to an AI product to deliver that daily interaction, and you don't see the difference between the two.

‍

Cameron (18:20.47)

Yeah, absolutely. It's pretty powerful. And you're absolutely right. And we really don't even seem to have the interfaces to take full advantages of the foundational models yet either. like as the...

‍

UX gets better. mean, ultimately that seems to be what you guys are doing, right? You're building a game, you're building custom crud and interface, like underneath all of it, it's a foundational model driving it, I would imagine. Maybe you refine it a bit, but nevertheless, like we're only now just figuring out how to like maximize utility out of these things. What do you think Elliot, having been in, you know, classic media, what underrated skill from television production translates perfectly into building an AI gaming studio?

‍

Elliot Wolf (19:17.241)

Well, I don't know if any of it's one-to-one. It's been a fun learning experience throughout, but I would say, and this is what I tell our writing staff all the time, the skill of writing is actually rewriting. And AI will allow you to solve for the blank page, but the...

‍

Elliot Wolf (19:44.385)

emotion that really makes these stories sing is going to come from the fine tuning. And I believe that that will become increasingly on the idea level where you sit down, you have a great idea, and then you'll leverage the technology to bring it to life. But frankly,

‍

I think it's the TV background and our writing staff and our creative team that makes Public Eye our first game sing and will make our product sing. It's instrumental in everything that we build having that creative eye and having the experience of telling stories for an audience and understanding how to create really compelling characters and create really compelling narratives.

‍

Cameron (20:40.824)

This is sort of an interesting question in that it's maybe counterintuitive to many people being super excited about AI, but it's actually, what do you think is the least realistic assumption that generative AI enthusiasts currently hold?

‍

Elliot Wolf (20:59.929)

That's a great question. I have a hard time answering that because AI has continually surprised me with how quickly it's improved. I believe that building original IP will always come from human creativity. Now, how you exercise that IP to what I said earlier in the conversation tomorrow.

‍

can be daily touches, games that extends the engagement and retention around the IP. Tomorrow, that can be UGC within that IP universe, where users are creating content or their own mini games that they're sharing with friends within that universe. And they're able to build within that universe because our platform is able to create those guardrails.

‍

with the IP guidelines, the narrative consistency, and so forth.

‍

that doubles down on this idea that it's still gonna be human first, right? Like when you get fans creating within their favorite IP universe, I don't think that any studio executive or even the creator of the IP is gonna be able to predict what they can make. It's enabling the audience and the people who love that IP the most to themselves be creative and AI is just a tool to do that.

‍

Cameron (22:40.312)

really interesting, a collective fan base influencing the corpus of the IP. Elliot, I'm curious, how might interactive AI storytelling reshape what we consider great narrative to that point?

‍

Elliot Wolf (22:57.665)

‍

I think the depth of exploration around the narrative will greatly change as AI becomes more mainstream and interacting with narratives, where today you create a TV show as an example. You go from script to screen and specifically your screen in the editing bay, right? And then you lock the cut.

‍

and it's distributed globally.

‍

In a future state, you can imagine where you create the universe and the underlying logic of the universe. And then there's what I would consider the next evolution of fan fiction, where people can explore that universe and uncover story beats and elements of that universe that would have never made it into the locked cut.

‍

that you are confined to with 60 minutes of runtime or whatever it may be. The exploration can be endless. And that can be in text form, going back and forth with LLMs who have ingested the initial creative from the creator or in a visual medium, a gaming medium, you name it. That's what excites me so much is six months ago, I would have said it was just text.

‍

Elliot Wolf (24:31.833)

And today I feel very confident in our video outputs. I don't necessarily, you know, I know that there's a lot of folks working on TV and films with AI generation. I think that that narrative consistency or sorry, that visual consistency, it's really up to the audience what they have the appetite for. But certainly in bite-sized pieces and individual clips that make up gameplay.

‍

It's very compelling.

‍

Cameron (25:03.756)

Right, I hadn't considered that for now, but like you think about you're playing a linear game or you're participating in the narrative as a player, you have an outcome, you know, pings VO3 or whatever it is. And then like, you can have a bumper and interstitial experiences to what just occurred during your game plan. That's, that's really a clever and awesome idea.

‍

Elliot Wolf (25:22.603)

You also have a video summary of the story that you pieced together. So let's use Public Guy as an example. You've put together what you believe happened and who did it. And if the cost of production is the cost of compute, you can get a personalized retelling of the story that you just put together, whether it be right or wrong, whether or not we've accounted for that, and then share that with your friends.

‍

Here's what I think happened, but it's video. And that's really compelling to me where you as the player are in the driver's seat with the...

‍

Elliot Wolf (26:09.249)

outputs from the game that you're seeing.

‍

Cameron (26:13.742)

pretty amazing. Well, tell us, tell our audience like where can they get their hands on this game? How do they follow you? What's the, what's the, what's the next steps here?

‍

Elliot Wolf (26:23.555)

So follow us on all socials with Wolf Games. We're at @WolfGames across social media. Public Eye is coming out later this year. And we have a ton of other exciting announcements in the pipeline of games and partners that we're working with. So stay tuned. But very excited to have folks start playing Public Eye and participating.

‍

Cameron (26:54.498)

Awesome, Elliot. Thanks for joining me on production and sharing with us just a little bit about your thinking about where AI is going, how it's going to intersect with storytelling. It's super fascinating and I know a lot of people will be eager to check out the gameplay and figure out how it sort of fits into their own workflows and their own IP.

‍

Elliot Wolf (27:12.697)

Thanks for having me.