"There will be AI tools and AI assistance in every craft going forward"
Markus Kiukkonen, COO and co-founder of Bitmagic, talks pivoting to AI game creation and the future of UGC.
Welcome to the latest edition of the AI Gamechangers newsletter, where each week, you’ll discover exclusive interviews with industry leaders. Today, we’re speaking with Markus Kiukkonen of the award-winning AI game creation platform Bitmagic.
Scroll to the end for the latest news and links, including Ludo.ai’s big update, Netflix planning to use generative AI to help develop its new games, and Oasis – no, not the band, but an AI model that generates games in real-time.
Please share this with any of your colleagues interested in AI's role in games, and check back next week when we will meet the Brimstone AI founders.
Markus Kiukkonen, Bitmagic
Today, we speak to the COO and co-founder of Finnish startup Bitmagic. The team recently won first prize in the Generative AI category at Game Changers 2025 (judged by a 22-person Silicon Valley jury). It’s developing an AI-powered game creation platform after pivoting from its original RPG concept in 2023. The platform enables users to create 3D games through simple text prompts in just 30 seconds.
Top takeaways in this interview:
Bitmagic pivoted from a niche RPG concept into a broader AI-powered game creation platform.
Their core product enables users to create games through text prompts, generating 3D worlds and quests in 30 seconds, with a focus on being more accessible than traditional platforms like Roblox.
The team predicts major changes in UGC over the next five years, arguing that existing platforms with "surface paint" AI will struggle against new companies built with AI at their core.
AI Gamechangers: Please tell us a little bit about your background in games.
Markus Kiukkonen: I started in mobile games when there were still Symbian mobiles and Nokia phones with premium games, like 99-cent text message games. Mr.Goodliving was a local company here. I've seen the good times and bad times in Helsinki’s gaming scene. I don't know if you remember Habbo from the early 2000s: I spent five years there, and it was fantastic. I'm really grateful for the opportunity. It was one of the first free-to-play platforms outside of Japan in the early 2000s, one of the pioneers of free-to-play services before people even knew [the phrase] “free-to-play”. So that was a really good head-start into that world.
Please tell us how Bitmagic came about – you made a big pivot to get to where the platform is now.
We started in early 2022. The company was originally called Roleverse, a “role-playing metaverse” in a sense. The idea was to build a multiplayer real-time role-playing game, where one of the players would be similar to a Dungeon Master or Game Master. It would be a group of players playing together in real time, and one of them would have a different view, where they could alter the game world with AI tools. The AI component was part of the vision from day one, but it was a very different flavour. It was more of a tool enabling one of the players to customise the game on the fly. It mixed the best parts of tabletop role-playing games and computer role-playing games.
“I don't see AI becoming a game designer from the point of view of creating experiences that are fundamentally intuitive for humans. At least for the foreseeable future.”
Markus Kiukkonen
But by the spring of 2023, the investor feedback was lukewarm at best. Everyone seems to feel like in order to make anything in the RPG space, you need a D&D licence, plus there are a lot of bad experiences with MMOs, and even if it's a small-scale multiplayer game, it sounds very complex. “It sounds kind of niche.” We started to realise that we wouldn't be able to raise a seed round with it, and we decided that we needed to move away.
But we had already started to build the AI tool. We realised that actually we could create more of a general UGC platform, allowing anyone to create games with these tools, not just the Game Master part. In summer 2023, we pivoted the company into what is now known as Bitmagic. Since then we've been solely focusing on building this platform where anyone can create games.
Can you tell us a little bit about the user experience and how that works in practice?
Our vision is that you start as a player with a text interface, and you describe what kind of game you would like to create – you describe the game in the text prompt. You hit a button, and in 30 seconds, you have the 3D world, the game experience, the quests, characters, missions, everything built – and you can play the game, or fine-tune the game further before sharing it with your friends.
Early on, it's only going to be third-person action games. But the long-term vision is that you could do games in different genres too. In a way, we’re trying to limit the scope somewhat, even though it's still a bit crazy, and then build from there outwards.
We see that it's the multiplayer experience that players crave, and thus, we're actually now building a multiplayer Steam Play test version, which focuses more on co-creation. We want to focus on a sandbox experience, where players come together in the same world, and every one of them can prompt and customise the game world and create a mini experience with emergent gameplay there.
AI itself is growing and evolving all the time. How have you stayed on top of that?
Early on [in our development], ChatGPT had not broken out in a large way yet. We were curious about the capabilities of large language models, and we were already thinking, “This will evolve. At any point in time, this is the worst version of AI tools we'll ever have. It’ll only go in one direction; it will only get better.”
We were very forward-leaning. We believed things would come. The development has been faster than anyone anticipated, especially on the cost-per-token side.
We benefit from the fact that we're a small and nimble team. We are constantly sculpting what's happening with the AI tool and what's happening on the white paper level and then trying to bring in new breakthroughs as quickly as possible within our own development process. There are big pros and cons because I have never seen technology moving as fast as it is now, in all my years of gaming and tech.
How do you make sure you balance the UGC creativity with ensuring the games are good quality and that they're safe? Do you apply guardrails?
[Look at] Roblox Studio – it’s fairly cumbersome to create games, and I think the quality suffers largely because it's so cumbersome. Our intuition, early on, was that we wanted to craft the controls for game-makers. How do you move the character? What's the responsiveness? What's the feel of the game? That’s not coming from AI; it's coming from decades of game-building experience. The controls and the feel of the character are what we build, and then the game worlds and the content is then mainly built by an AI.
“Every single profession will start to look different. There will be AI tools and AI assistance in every craft going forward. You have to embrace it to keep up with the rest.”
Markus Kiukkonen
I don't see AI becoming a game designer from the point of view of creating experiences that are fundamentally intuitive for humans. At least for the foreseeable future.
What about things like copyright? We want to understand some of the ethical issues with AI. Does the role of the player in creation change any of the legal terms and conditions?
We want to offer as much expression power to the creator as we can and, at the same time, respect the copyrights. This could mean, for example, that you can create games for personal usage with no limits, but you would not be able to share or monetise these experiences if any copyright violations were found. In the long term, we want to partner with IP holders so that their IP can be used in the platform and the IP holders can get part of the revenue. A little bit like what YouTube is doing with music.
We do want to look into partnering with the creators, similar to Roblox and Fortnite – having an actual creator economy where we share the revenue with the creators. So that's the current thinking around that but again, we're still not really public yet. It’s early for us.
You recently raised a funding round. What will you do with your investment?
It's really scaling up the dev team. There’s a tremendous amount of things we would want to do, but we’re short on hands. It's mainly hiring more engineering power! It's not necessarily a lack of experts (the core team covers all that really well – the understanding of AI and the game-building and how to merge those two). It's more about throughput and having more capabilities to build more things simultaneously. That's mainly where the focus is, and the funding allows us to move faster.
And in other big news, Bitmagic just won first prize in the Generative AI category of Game Changers 2025, hosted by Lightspeed, GamesBeat, and NASDAQ! Congratulations. Can you tell us about that?
This is a significant achievement for us. The 22-person jury included some very influential people from Silicon Valley, so having them decide that Bitmagic is the standout winner in generative AI is massive news for us. We are very grateful for the acknowledgement, and this gives us new energy to keep pushing the boundaries of what is possible with AI in gaming.
In October, you spoke at the PG Connects Helsinki conference. Your talk was called Strategies for Building the Tech that Powers Today's Top Titles. Can you share any insights from that?
There was a time, at least in Finland, when people thought you needed to build everything by yourself. That was because maybe third-party tools were not advanced enough (or there wasn't availability). But there was also a feeling that it doesn't look too complicated. “We could easily do it ourselves!” The trap people fell into is that it may be relatively easy to do at first, but you need to maintain it; over time, the third-party tools (built by companies who solely focus on those tools) have a competitive advantage over your in-house tech. It's a balancing act.
My point in that conference talk was identifying what's crucial for whatever you're doing. What is your core unique selling point? How are you trying to differentiate yourself? It was a varied panel, so there were different perspectives, especially from the perspective of a young startup. How much can you use external tools to just speed up things, so that you start to learn with your target audience? Is this going to fly? And there, I would really think twice about what you build in-house and what you buy from external sources. Of course, if you are a large, established company, then it's a very different situation. You could have in-house tech teams. But for young companies, I think third-party tools are better than they’ve ever been.
At Bitmagic, are you developing your own internal models, or are you using third-party tools in your development?
We extensively use third-party solutions. Again, we're not building foundational language models for ourselves. I don't think anyone can compete with the big boys. We take whatever is the latest, greatest models from whoever happens to be leading in the race.
And we're [also] not building multiplayer tech from scratch. We'd rather lean into others’ expertise because we want to focus on how we can best showcase the AI's capabilities to build games. We want the team to focus on the very heart of the experience.
How do you imagine the role of professional game developers will change as AI becomes more prevalent in the industry? Do you think that as a profession, it will have to adapt?
The short answer is, “Yes!” I think every single profession will start to look different. There will be AI tools and AI assistance in every craft going forward.
From my point of view, you have to embrace it, in order to keep up with the rest. If someone is using AI tools and they're 30% or 50% faster than you, there's no excuse for you to be slower. I don't see that there’s an option not to lean into it. It's just a question of how to manage change and how to get the organisation behind that change. In large organisations, I can imagine that, like any kind of major change, it can be painful. But every single craft is going to use AI.
As with mobile phones, for instance, or any other technological innovation, it also creates a lot of new competencies and new roles. For example, in mobile gaming, free-to-play games brought a lot of product managers and it became an economic specialism. All kinds of roles came with mobile games. I don't know if any company has actually hired professional “prompt engineers” – it was a word, maybe a few months ago, although I don't know if it's going to be an actual profession (I think it's more likely there will be craft-specific tools).
You still have to have the domain expertise, and that's why I believe that your craftsmanship and experience still has value. It’s just that tools will allow people to create significantly faster, so the cost of any kind of content is going to keep going down. That, of course, is a concern in itself. There are a lot of implications from the fact that if the cost of creating almost anything starts to head towards zero, then that world will look very different from what we have now.
As you continue developing Bitmagic, looking ahead five years, what do you predict will be the most significant developments in AI and UGC?
Things are moving faster than anyone anticipated, so it's hard to see that far in the future. A big part of why we are so excited about Bitmagic is I believe that a lot of the existing UGC platforms have major disruption ahead of them. Old dogs rarely manage to navigate the challenges of a massive paradigm shift. It's mainly because of the size of the organisations and their existing very profitable businesses that they don't want to disrupt.
I would imagine that if we look at the top five UGC platforms in the world now and go five years into the future, that list of the top five will look very different. Maybe there are still some existing companies there, but there will definitely be new ones. That's a tremendous opportunity for us to create a UGC platform that has AI built into the core of the experience. A lot of the previous generation of companies now have a “surface paint” of AI on top, but they still have their legacy engine, and I don't think it's going to be very easy to get the AI at the heart of the experience without rebuilding it from scratch. That will leave a lot of room for newcomers – that’s us – so I think it's going to be very disruptive, for sure.
Further down the rabbit hole
Your handy digest of what’s new in the world of games and AI this week:
Former poker AI researcher Noam Brown revealed how giving a model 20 seconds more “thinking time” was worth 100,000 times more training data. Could slower, smarter processing revolutionise game intelligence?
Ludo.ai has evolved into a complete game design hub with a major update this month.
Microsoft Xbox is rolling out its new AI support chatbot (“Support Virtual Agent”) to help players solve console and game issues.
EA Sports College Football 25, 2024's top-selling game, added Adobe Express in its update last month. This means players can design custom logos for stadiums and jerseys using the tool’s generative AI capabilities if they wish.
Netflix plans to use generative AI to help develop its new games. The initiative will be led by games industry veteran Mike Verdu, who posted on LinkedIn that this is “a ‘once in a generation’ inflection point for game development and player experiences”.
Decart and AI chip-maker Etched unveiled Oasis, an AI model that generates playable video games in real-time, without a traditional game engine.
