Panel: All In Or Incremental? What's The Best Way To Integrate AI In Your Production?

Okay,
cool. Okay. So, uh, welcome to the stage, the next panel. Um, all or incremental. What's the best way to integrate AI into your production?
Hello. I'm having to stand because we can't fit enough chairs on the on the stage. This is a bit bit uh strange, but um, thank you so much for joining us um, this afternoon to discuss, yes, should we be going allin or incremental with AI adoption in our production? um timelines. Um we met last week to sort

of discuss this ahead of the panel. Um there's a lot to cover. It's a really broad topic, so we're going to try and sort of shoot through it um as best we can um and cover as much as we can. Um so bear with us. Um but first of all, it'd be great if you could all please just introduce yourselves um where you're working and then also um just any any summarize sort of briefly how you're utilizing AI in your current workflows. That would be fantastic. All right, I'll start. Um, so my name is Jaffet Asher. Um, I have my own company, Polarity Reversal Limited. Um, I feel like a bit of a fraud being here because I'm not really a game maker, per se. Um, though I do make interactive fiction and have explored game experiences in a number of environments. Um, I'm probably best known as the creator of Liquid Television that spun out Beaver and Butad Neon Flux. worked at uh worked at BBC Children's running all of their interactive games, interactive episodes,

digital um and then got involved in AR and most recently have worked for the last couple of years in Matters AI studio creating uh characters, AI characters, exploring real-time AI and how that can be used in storytelling and indeed in game play. So that's the intro.
Hi everyone, my name is James Wu. Um I'm the general manager of AI partnerships at uh gamebot. So I've been working in the game industry for more than two decades and uh uh span at EA uh Tencent Nvidia like by dance games. Um also uh I had a two startups uh during them we tried to build the gaming consoles in China for China market and also try to challenge Xbox and PlayStation. I know that sounds really crazy, huh? Uh, and it turned out uh to be not a big success for me. Uh, obviously.

So, at Gamebot, uh, at Gamebot, we uh help, uh, developers and publishers. Uh, we, uh, deliver AI agent solutions. Um, our key customers including MLBB, uh, League of Legends, Wild Rift, uh, PUBG Mobile, and Garina Free Fire. Um we deliver measurable results. Um we have track record that uh our bots can help to boost the user retention up to 6% while the uh monthly active user up to 15%. So personally I'm a hardcore gamer and sports lover. Anyone of you loves the Monster Hunter series or the Liverpool Football Club, we could be really best friends. Thank you. Hi everyone. Uh I'm he I'm the CEO and co-founder of IBA uh out from Norway. So very happy to be here. It's not 2 mters with snow here.

So that makes me really happy. Um I'm running a company where we use AI to make the communities in games and other platforms safe. So we have made a moderation platform to detect the bad uh players or the bad behavior before uh the bad stuff happens. So we have been focusing on uh protecting kids and teens. Uh but we see that our technology can be used for uh other kids like uh the older kids like me because we we are all uh when we are in the gaming world we get to be a little bit enthusiastic as we can if we can put it that way and someone needs to uh be heard sometimes that maybe you should tone yourself down a bit. Um, we are building uh AI in the company, but we're also using all kind of AI out there to make us uh to make

the company run uh faster, more efficient, etc. So, u happy to be here and uh I'm not a Liverpool fan, but uh yeah, hey, we have a Norwegian soccer player, so he's really great. Hold on. Yeah.
Uh, hi, I'm David Himeneth, director and co-founder at Two Awesome Studio. We are a 30 people development studio from sunny Barcelona. Um, I guess the weather goes from Norway to London to Barcelona. Um, we develop our own IP. Uh, our next game is Alterma coming out uh this year for PC and consoles. We did Axelottal before that. uh we also do a lot of code development and porting for uh publishers and and studios. So I bring the vision of uh more of a traditional studio on development and and how and the different ai tools are are used by our teammates and I'll I'll bring

opinions from my team's uh members in in that respect.
Hey, I'm Vulcan Girl. I'm the co-founder CEO of Layer. uh we're building the operating system uh for game studios to use AI. Our platform is aims to be the kind of single solution you can use for all sorts of content generation 2D video 3D audio and uh we've been working for the last three years with studio many studios and we've seen kind of like the spectrum of kind of the level and commitment of adoption. Some have gone all in from the beginning and some have been a little bit more reluctant and hopefully we'll be able to share some insights on kind of how those different approaches resulted in different outcomes. Uh so not from a game developers perspective but from a tool provider perspective we got to see a wide range of outcomes and approaches. So hopefully that'll be useful to share.
Great. Thank you. So I think obviously across the board like as a litmus check um AI has been adopted and or or very

much our core focus um and I guess sort of in a current reality check and sort of like what's happening versus the the hype around this um are we talking about sort of tool level improvements here or like fundamental workflow change? Where do you think we are at this stage currently right now with your work and your your your customers work? U I I mean everybody comes at it from such a different point of view in this panel I suspect. Um for me um I I'm thinking about it from a storytelling point of view and from a development of uh uh story worlds. Um and I also think about it in terms of real-time animation versus animation as a tool within a pipeline. And I think they're very different things. It's important that people understand they're very different things and integrating them into your workflow. um is really different depending on which particular approach you take. And what I mean by that for those who aren't familiar is that if

you're working with real-time animation as a creator, you're setting up an environment for a user to experience AI directly through conversation with it in directions that aren't entirely going to be in your control as a creator. You can shape and frame, but you know, there's going to be improvisation within that. um versus using AI to generate outcomes which you curate and only when you're satisfied with your curation would you put in front of a a user and they're quite different things and they have quite different risks um and rewards. I I always say to people that if you're not ready to try using AI in your pipelines now, then you probably never will be because you're making those decisions based on not the state of the technology, but in your view about AI and generative AI as a medium and as a tool and um it's not perfect for everything. It's great for experimenting. You have to

retain control of it, use it to amplify what you can do. Um, so that that's sort of the main stuff. And then finally, I would just say that for me, it's really changed the way I think about building a story world and then exploring within it because AI allows you to do that in a way that's different from the way it used to be. Um, and you can choose to open that up to potential users if you want or potential funders if you want or you can carve out pieces of narrative that you discover in that process for different platforms and different environments. So on a very core level, it's changed the nature of how we develop or how I develop IP and think about IP and explore my own ideas. Yeah. And uh I can also just follow up like AI is not the answer for everything uh obviously and also trying to figure out what is the problem you want to solve and how can you best solve it. Uh

and we also see that uh if you have have a problem to solve uh you also need to have control of the data uh unless if you don't have the data and you don't have uh and you don't know what your data means. Uh well you you should not use AI because it will give you the wrong answers obviously.
Yeah. Bringing more the uh development studio perspective into into the mix. I agree with uh what you're saying that it's regarding your question. It's for the current state of affairs is good for certain things is not there yet and I don't know if it will ever be there yet for some other things and and we see that especially in different disciplines of of development uh and we can get into that uh a bit more later and I'm disregarding for the moment the opinions ethical moral dilemas that uh some some

disciplines may have about it uh but from a purely productions perspective uh point of view it is ready and useful for uh some areas of development and I think it's more of a hype and not yet ready for some other things
and I also encourage everyone to think about this as like there's say productivity improvements you can have today with you know AI for example like oh I have an art pipeline I used to do this thing manually now I can automate it that's the I guess the first step, but also think of it as this trend that keeps improving. I was just having a conversation with today with someone like the beginning of say 2025. It just seems so long ago where like the video models were really bad and the image models were kind of getting there good but not as good as now. And then fast forward a year, it seems like a decade has passed in terms of quality, speed, performance. And then think of this trend kind of going forward. uh there's

a lot of stuff that we believe will uh be possible only with AI things like personalization, customization, culturalization where unless the team has adopted the the tools, then they're going to keep kind of falling further and further behind because they will not have that foundation to build upon. So it might seem painful at first for teams to adopt and switch. And I I would argue there's value already, like massive value we're seeing with, you know, what people are doing today. But it's also the the future opportunities that people would be missing out if they don't get their teams kind of on boarded and and adopting this new technology.
I think that's super interesting. I think that's one thing that people quickly forget is how much new kinds of game narrative experience are being opened up. Um, and so if you only focus on how you're going to replace parts of a process that you're already doing, you're not really considering the opportunity, the potential and the

growth for your studio, for your people, um, and for your customers. Um, and that's the exciting bit. Yeah.
Yeah. Something I want to sh you know share is that back in China the almost every boss in the gaming company they encouraged their team to embrace AI even in the toilets there were like labels on use AI embrace AI right but after they have set up so many invest so many human resource set up their AI teams they don't really know what to do
that's the reality right now
I'm sure I know what to do with AI in the toilet I I want to But yeah, that's why we're here today.
And I think that kind of leans into so I think we're all sort of leaning more on the side of incremental rather than allin given the state of the market today. Um but is it is it a tech decision or is it a change management decision in an organization or is it is it both like you know I think it it

tends to always fall to someone in a tech position but has that shifted like you say even since the beginning of of last year? So I maybe I would actually argue going all in. We've seen a couple companies like some large ones do that. uh and uh you know it took a while but now we're seeing they're getting significant uh advantages compared to some other studios that have kind of gone incremental and it's not easy you know to it really has to sometimes like not just the co saying top down I want more AI but actually doing things like hiring a team having that team be responsible for certain outcomes and measuring those outcomes and really taking it seriously isn't just about saying I want my teams to adopt AI but really investing ing in processes, changes, giving the team space, giving them a little bit of a like a room to say, "Oh, you know what? While we're exploring these things, there's going to be some gaps and other things we would have been able to do, but we're adopting this new process, new technology." So, it's going to be a little bit of a

period of say not as much stuff we can do in the other areas, but once you get over that hump, they're just flying now. Like there's some studios with hundreds of artists using our tool and they are now very very happy because they got over that hump and now they're like they went all in in the beginning and now once over the hump they're extremely happy whereas some other ones they're like they gave their teams a lot of freedom to go and explore tools. There wasn't like a strategy top down. So that kind of really misses out on a lot of the potential. I mean again in our case not talking about like a product integration but an internal tool. A lot of value comes from not just one team adopting it but like having like a central repository of models you share fine-tuned models. um certain shared knowledge across the team on that tool like adopting a single solution that can actually serve the entire team of hundreds of artists. Not just artists but product managers um you know other other roles that are not really art focused also were able to become really

productive uh with a top- down adoption of like a single solution. So we saw that like all-in approach really pay off for a lot of customers.
Yeah. And also following up on that uh it's an it's your assistant it and also it cannot live by itself for now. It needs a human I believe to been looked after or else you can uh end up in some really bad uh front pages in the newspapers like we've seen on Grock for example. So I'm 100% on your side like yes deep dive into it but also make sure that you you are on top of it uh that you have control both on the data but also on the decisions and how it uh moves forward or else you might been you might have some issues.
I'm going to to bring a bit of a counter argument to to to what you are saying.

Um I think it's important as well to see what's the status of your pipeline your projects and what you have ahead because if you are let's say in the middle of a production which is already running for six months to a year and you have one more year ahead of you disrupting that workflow and saying okay we're going to go all in into AI now we're gonna destroy the foundation and rebuild it with AI I with an AI pipeline, it's going to cause not only technical friction, but it's going to cause a lot of human friction in your team. And when you may be saving in the long term, you're going to be losing it just by doing that shift. It's is it's almost like changing to a different engine in the middle of a production. Well, it can be done, but it's going to be painful. So my advice is that there's no one sizefits all um from running

our studio. We have seen that in some areas people really clearly saw okay I can accelerate what I do without having to rebuild everything from scratch and I'm happy to do it and that's that's what we said okay it it makes sense to do it there. Would we do it the same way if you are starting a company or you are starting a a game from from scratch? Well, that's that's more I guess where you come to your question and then it's okay. I I want to build it as an I first uh project and it may make sense. Um in the middle of a production it's a bit more difficult. I would say
yeah I think that's really interesting what you said. It's like uh I I think this is a false distinction incremental or allin you know it's about using tools in appropriate ways for appropriate tasks and that's always true and that's as true with AI as it is with anything else. So yeah
and you mentioned um sort of hundreds of artists and you know do you think it's a

a level playing field across the industry? So if you look at like mobile versus PC console, indie developer versus sort of AAA um
I is it is it level or is it not because you know budgets aside just the to David's point you know if you're midway through a production cycle y
that's really challenging. So
absolutely and plus not even across the industry across different companies but within a company there's different levels of adoption. So I'd say I think you're absolutely right like maybe that distinction doesn't quite make sense. It's more about uh do you have a strategy rather than like a high level oh we should use AI it should be a little bit more deliberate and uh if you leave it to individuals or individual teams even it becomes a little bit kind of like missing the point and missing the real value it should be kind of driven with a strategy rather than kind of like let's say fair let every kind of team handle it on their own type of thing it's it does make sense I mean I'm talking like large organizations of

course if a small team you just talked the person next to you and you can figure out something. Uh but to the point about production timelines, um you're right, it is a disruption and um you know, you can kind of choose which way to go, but we were able to work with existing art pipelines with hundreds of artists working on them. And they were kind of I mean, you're always in the middle of something, right? There's never like a free moment in, you know, there's always the next game you're preparing for. There's never like this, okay, now we have a month off. We can do uh anything we want. So it it really does take kind of someone willing to take that pain to get over that hump. But once you get over the hump, I guess my point is there's a lot of value there, especially if you have a year ahead of you to get to a certain point, that's plenty of time to reap uh the results of that choice. If it's like, oh, I'm trying to get to a deadline in a month, maybe it doesn't make sense to disrupt things. But, you know, it really uh it's something that, as you said, teams have to decide on their own. But I should

sorry what the other thing I want to say is within even a large company a lot of say UA actually gets a lot more value on on certain cases than say game production. Um so even within a team within a large company across different teams it's not a level playing field.
And I guess um just conscious of time moving on to kind of ethics and sort of player perception around this. Um there's been some quite big headlines as you say, people very much all in and then some people who are very, you know, firmly they won't accept AI generation or AI um even placeholder art in their sort of greenlighting processes. Um how much do you think players should know about AI usage in the games they're buying and playing and how do we communicate and not lose trust with our customers and and players?
I I can quickly jump in. Sorry, I I was the last speaker. I'll just say something quick which is I I I see it as um I think there's a lot of fear and hype for various reasons. Um but I think

it's it will be similar to someone in the early days saying oh this game is made with Photoshop like oh you know what's the big deal about that like right now we don't even care. I think there's going to come a point in time when it's going to become a natural part of the game development cycle. I personally don't believe it's going to replace um jobs. I think it's going to replace tasks and jobs will evolve. Again, that's a whole other conversation, but I just want to to just say I don't think like those ethical concerns will be that long lasting. There's a lot of things of course that you know data protection, privacy, IP rights, those have to be addressed with, you know, tools that for example we address them ourselves already and we try to make it easy for people to be compliant. Uh, but I don't think there's really a huge need to say this is like an AI generated content in the game similar to you wouldn't say this is made with Photoshop or Maya.
Well, I think uh first the key is to be

100% transparent because um if you use AI uh for example in gamebot uh our solution uh we have invisible AIs versus visible AIs. When we use visible AI, usually they're AI boss or AI co-pilots. Not people knows that, right? And let's well what what it goes to the invisible AIs. It's just like game bots that play with you. Maybe you know some of your teammates or your opponents. They just, you know, we use our deep learning AI bots to replace those uh behavior tree AIs uh to look smarter, to look like a more like a human being. uh 99.5% of our bots won't be fi find out to be AIS. So um it's just that we use a modernized technology to provide the gamers a better experience. It's one part of the game feature. So right uh it's always being transparent but while

sometimes you know the gamers doesn't have to know.
Yeah. And I I want to follow up on this because we also see there are some AI boyfriends, AI girlfriends out there and uh even though the uh the kids or the teens or the adults, they know this, they still get connected. So I think it's also if we're looking into the ethical uh space, we should also develop uh services that we act that doesn't actually turn into something we don't want them to turn into. So as humans I think we need to be uh more on top of the development more than sitting behind and thinking this is happening. Yeah, this is happening. You have to be there.
Yeah, there are three different pieces to it, aren't there? I mean I think I understand people's concerns about particularly copyright. We all understand this as creators. I've had stuff of mine ingested into these models, but then I've also contributed to the models by writing golden sets for

them. um you know uh but I and I hope that eventually there will be some kind of reckoning around that because there needs to be but the door is kind of bolt you know the horse is bolted that you can't overcome that um and I personally believe that you know if you're a creator you you move forward you don't constantly look back at what you've done before and this is how I've always operated anyway um I do think that the issue of Um, suspension of disbelief is something that we do as storytellers. Um, it's a it's a large part of our our goal is to get you to stop thinking about the strings and engage in the narrative. And the temptation to do that with Gen AI in real time is very high, but obviously it comes with massive emotional, moral, ethical risk and particularly with young audiences. So, you know, in the game space, in the

role- playinging space, we have to be extremely careful about signposting. And I think there's a societal obligation to think this stuff through, too. Um, this is what I think Kagan was starting to talk about. I think, um, that's a big big concern of mine personally. Um, and you know, uh, AI literacy, critical thinking, these are the important skills of the future in an environment where you can trigger stuff. And the final thing I would say about this is that for me anyway, AI does AI is does nothing. AI has no objective. We ask it to do things by prompting it whether that's in a conversation or an instruction and then it does things and we then decide whether we like those things or not to use to consume to think about and go and use and ask it to make something else. So, it's really important to remember that in terms of the how it works and the ethics of it and how it's incorporated into artwork because

somebody some human has made a decision to trigger a set of computational actions and then said, "Yeah, that's the computational output I was looking for. I'm now going to place that in my context artistically." Yeah.
Yeah. I agree. there's always a human driving it.
Yeah.
Um I think one of the points that makes this uh that that drives the whole ethics debate around this is is is actually money. Uh you were saying that uh it's not the technology that is supposed to replace jobs but the fear is there and it's very it's a very valid fear and many people won't change their minds even if you tell them no it's going to help you do your job better or whatever. We are seeing constantly more and more jobs getting cut in the game industry and and we are in a very difficult situation at the moment in in the industry. So this adds fuel to the fire. So it's it's not easy to to tell people now you're going to be using a tool that is going to be learning what you do. It's going to be trained on your

data and the more it becomes better. Maybe not you but your colleague is not needed anymore. And and let's let's be real that's that's happening. and uh in some fields more than in others and it's more uh it has a more direct impact in in some areas of of development than than in others. So I can I can relate. I I have been having uh very deep conversations with uh team members and and also with uh people outside the studio that uh that are in in in the field and what's what are their views on this and whether we like it or not AI is for the moment is here to stay unless there's a big bubble burst and and it all comes down but the way things look uh it's it's not going to go go down and we will have to to see and coexist and adapt with it That doesn't mean that it's that the push back is going to disappear. It's going to continue and it's going to continue because it's a

very valid reason is money and people want jobs and want to keep them. And this yeah it will create new jobs but it's also destroying some of of the jobs in in the market. It's a disruptor like any any new technology that okay you create new jobs based on that technology and some jobs become obsolete and it's it has happened many many times over the years in our industry and in others. Um from the consumer point of view I see something a bit different because I think it's more generational. If I ask my mother about the Coca-Cola announcement she wouldn't even realize that it's AI. It's just shitty and that's it. um but she wouldn't even have an opinion about AI or how it was made or whatever. And then you have my kids the and again that generation doesn't care because they see it more as a b or a or a or even a style. uh it's it's more our generation in the 30s4s

where we see the the the difference between okay what was made by traditional art and what was made with GNAI tools and and there's a big gap at the moment and and we see that we see the the link to money and to ethics and to and to all of that and then there's that push back so that's what we see um yeah we all have to adapt
yeah and I think as an industry maybe like you say, looking to what other industries are doing. I think we can be quite bad at that sometimes. We kind of quite insular and um worry about our own thing and don't look to sort of the way other other industries are maybe tackling creative industries are tackling the problem themselves. Um we're at time. So just super quickly if I could just get sort of a lightning close um from each of you on kind of what you would say a studio should do about AI in the next 12 months if if you know they're kind of at this sort of dipping their toe in at the at the minute. You know, what would be one thing you would advise them to do? start that any work to say that?
Sure. I mean, I I'd say I'm sure

everyone has tried something and and you know, it's not like I I don't think anyone here has never seen what these tools can do. I'd say um if you have concerns on the team or whatever the blockers are, just know that there's teams out there that have really resolved those blockers and um I would encourage you to really kind of adopt a strategy around this. a little bit of a kind of companywide strategy and really go after it. I'm I'm I think we there when we say AI, by the way, I think there's a lot of different ways to use AI like inside the product or as an internal tool. I'm specifically referring to it as an internal tool. Uh and I think as an internal tool, there's a lot of success stories. there's a lot of value people already got and there's just like a trend towards uh very differentiated ways to use AI and that are I believe are coming very soon that teams who haven't adopted it will will just fall further and further behind. So my advice would be really try to kind of

adopt a strategy and go after it and uh do it kind of as a companywide thing to fully realize the potential across the whole team. Uh my point of view is a bit more nuanced. This more not adopt but more adapt. Uh use it where you see that you can use it where where you can check as I said the technological factor if it benefits your production but also the human factor are these changes going to mean that maybe you see 10% of your top talent walking out the door because of ethical reasons. That's al also important to consider in the middle of a production. So be a bit careful how you manage and check your team. See how they see these tools, where they see the value and then if you see areas where you can adapt to it and use it properly, uh yeah, go all in those areas. Yeah, following up on that, if you're not there already, please start using

it, but also with care. And uh if you have a friend or two that or someone you look up to which are really good at using it, ask them. And I know that everyone is so happy to provide insights into how they're using AI. So if you're interested in how we do it, happy to have a follow up afterwards. We're also up in the boot area in the startup zone in the in the blue room. So yeah.
So I think uh either all in or incremental I think for strategy wise you should all in uh you should embrace AI use it as as many as possible but uh tactical wise you need to go step by step and integrate AI into your business in a smart way. uh maybe just leave the leave the uh experts to do their own job and um maybe um just you know work with the experts and grow your team. I think that's uh something I see like some of

the smart uh companies back in China what they're doing right
for me I would encourage you to lean into the weird recognize that AI allows you to do some impossible things. a natively surreal medium and therefore you can tell new kinds of stories. And if you really don't know what to do with that idea, um, come talk to me. I've got lots of weird ideas. Um, seriously, it's a chance to take things in new directions and explore rather than just trying to figure out how to make it quicker or cheaper to do what we've already done.
Brilliant. Thank you all so much for your expertise and your insights and um, yeah, I think it's clear it's it's here to stay. Um, and it's how we adopt it that matters most. So, thank you. Thank you all very much.