Interview: SEED CCO on AI Tech Use, and Why Players Love Dictatorships

Video games have been trying to replicate entire universes since the early days, with pilots of the original Elite seeking out life in other galaxies all the way back in 1984. Technology has improved a great deal over the past forty years, allowing the developers of the upcoming SEED to create a living universe, one that never sleeps.

SEED is an MMO that never switches off. The player creates a Seedling (think The Sims, with the graphics replaced by Runescape characters) and can join or create a society on the planet Avesta. With the player's direction, the Seedlings can build and craft new settlements, form governments, create businesses, or even fall in love and have kids. Players can even communicate with their Seedlings using AI, which will respond to questions and comments.

DualShockers recently attended a preview event for SEED, which is set to enter early access in Summer 2026 and is coming to Steam in Fall 2026. It was here we spoke to Klang Games Co-Founder and SEED CCO Ívar Emilsson about how the game was created, the use of AI tech, his response to fans who criticize AI use in games, and why dictatorships are so popular among the fanbase.

Interview with Klang Games CCO Ívar Emilsson

Q: Hi, this is Scott Baird from DualShockers, speaking to…

Ívar Emilsson. My role is Chief Creative Officer and I'm also one of the co-founders of Klang Games. We've been at this now for about 12-13 years, so it's been a while.

Q: What inspired the creation of SEED?

Originally, I think it came up as an idea while me and Oddur were working on EVE Online. I don't know how familiar your readers are with EVE Online, but the learning curve is quite steep. Getting into the game is quite a challenge, but there are certain things about that virtual world that are absolutely amazing and groundbreaking.

The way that communities can form and people can create these social contracts and actually stick in the same persistent world for years was something that we saw as the future of gaming, I guess. The only problem was that I didn't remember I was working on the NPE of EVE Online at the time, the new player experience, and setting up all of the tutorial steps and the funnel and whatnot. Just the sheer amount of people that actually did not make it into the game, it kind of felt like a shame that they were missing out on something bigger that you could experience through this online community.

That's kind of where it started. We started thinking, how can we make something that is as grand and as deep and requires as much collaboration as EVE Online, but has a lower barrier of entry and will allow for a more broad audience? That's the absolute beginning, I think.

Q: Why go for a simulation MMO over a more standard combat-based one?

That also actually comes from one of the design issues that we believe most MMOs have.

It's kind of like the empty world problem, where in games like World of Warcraft and EVE Online, when you log out of the game, you're kind of gone. You're removed from the world for a certain period of time. What really struck us was the mix between persistence and simulation, so having persistent simulation.

You can think of it like a Tamagotchi, for example. It's a very simple simulation, but it's persistent. If you reset it, you're killing your Tamagotchi.

That simulation, that it is persistent, actually makes it really meaningful. There's a reason why people feed their Tamagotchis every day, because they have built some kind of long-term want with it over time.

One thing that's really funny with massive multiplayer games and just social networks in general, is that we've kind of moved away from being session-based in these things towards kind of blurring the lines between real life and the game. If you remember back in the day when you were on MSN or something and you would log out, or your friend would log out, they'd be greyed out. You can't even send them a message.

They have to be online for you to actually send them a message. That's kind of how still most MMOs work. The idea with SEED of creating this persistence was kind of blurring that line.

It's very similar to what you see in today's social networks like Instagram or Facebook. You don't care if another person you know is online or offline. You send them a message and you expect them to actually respond to it when they come back to it.

I think a simulation game is also great for setting up stuff where you come back to a changed world state. You as a player, you set something in motion. Maybe you are building a house or you are building a relationship and you come back the day after and something has actually changed. I think that is really fascinating. It's one of the things that we were trying to go for.

Q: What role do you think AI plays in the game going forward, especially with the tech improving so quickly?

Honestly, I wouldn't say.

I don't know. I do believe that there is a lot of clever stuff we can do with LLMs to make Seedlings more human. Just to make it clear, SEED is a human simulation.

It's not supposed to be the perfect simulation. We are not trying to create some robots that know the best way to X or Y. We are trying to capture the nuances of what it is to feel a certain way, choose something that might not be optimal, or be ignorant about something, and make a decision you might regret later. We are not doing machine learning.

We are not trying to optimize some algorithm that makes the perfect human. We are trying to simulate human beings. We are very imperfect. I think the imperfection is something that we want to lean into. The LLMs right now are really great at filling in the colours of all of the mechanical lines that we set with our utility AI. A utility AI purely tells the character what is best for them to do in this certain situation based on math.

But then, how do they interpret that? Why is this the most important thing for them? Being able to colour that with an LLM is really fascinating and truly brings it to life. Another thing that I will say is that originally, when we were thinking about seeking-to-seeking conversations and even seeking-to-player conversations, we were just thinking about abstracting all of this out. We will just have icons like The Sims does it, so people can read in between the lines and make up their own narrative.

Because anything that we did, any non-LLM technology solution that we tried to utilize, just wasn't there. You kept seeing repeated conversations or something that just didn't fit. Our monkey brains are crazy at recognizing patterns.

So we figured that it would make much more sense to just make it visual instead of text-based. There was just a moment in time when GPT 2.5 or something came out and we were experimenting with this stuff. We were like, okay, this actually feels a little bit weird.

It was almost uncanny how close to this being a real thought that a seeking could produce or a real sequence of events. I think we've been really in the whole LLM sphere for a while. We've been following it a lot.

Obviously, we're not doing anything crazy. We're not generating 3D models or any crazy assets or anything like that. Currently, we're just using it for speech between Seedlings, speech between the Seedling and the player.

We're using it for, I think, some icon generation where we've set some certain parameters when you're creating a corporation. And then, obviously, we're using it in the behaviors to colour on top of the utility AI. To answer where we take AI or LLMs or generative stuff from here is very hard to say.

We obviously still really want to create a unique art style. We have really strong art direction and really strong artists working with us. That isn't where we're stretching the things, but we're more stretching it into how can we make genuine virtual human beings, if that makes sense.

Q: There's quite a vocal backlash to AI systems in gaming online. What would you say to anyone who would dismiss SEED outright due to its AI content?

I think it really depends on how you use it and what you're using. Obviously, if we're using some unethical models or we're using something that is stealing people's work, I can definitely understand the sentiment.

Personally, to me, I feel like when you are using it for genuinely things that may improve the player experience, and in this case, it's creating a bigger emotional attachment between you and your virtual character, I think that is... Personally, I think that is fine. I understand the criticism, obviously. But the fact is things are changing, games are changing, entertainment is changing.

We'll do our best to keep it grounded, keep it creative and driven by real artists. I don't know if that was a super political answer or what. Obviously, anyone can have their opinion on anything.

I sometimes think it's sad that people put everything into the same category. A game that completely AI generates all of its art, UX, UI and interactions. I don't know.

I haven't seen anything good yet. Maybe there will be something eventually, I'm not sure.

Q: I noticed in my Preview of SEED that there's a lot of steps to every action when a player does something. For example, if you want to take wood, you need to chop down the tree, then pick up the wood, then take it back. There are a lot of steps to complete something. Are there any plans to streamline that in the future? Or is that the intended vision for the game?

One thing that we believe matters a lot in massive multiplayer games is creating different roles for people to take on and feel like they are valued members of the group that they're in.

This is a little bit like a game design speak. Creating roles where people can optimize by spreading out responsibilities is a design pattern that we are pursuing. You can be a jack of all trades, but master of none.

Or you can really focus in on a certain thing. In order to produce the best results, you have different Seedlings fulfilling different responsibilities. In this case, you might have a seedling that is really strong and is chopping down trees all day.

Then you might have somebody that delivers stuff from A to B, and you set those two things up together in order to create the better result. That is player-to-player coordination and communication.

Q: Are there any plans to bring SEED to consoles in the future, or is that an impossibility?

I think it's definitely possible.

We are seeing more and more games on consoles that normally you wouldn't think would be played there. Currently, we are primarily focusing on PC, Mac, and Mobile. Mobile being mostly a companion app at this stage.

The current camera controls sometimes struggle to follow the players and the Seedlings. Are there any plans to address this in the future?

Absolutely. What you looked at today is obviously our development build.

It's actually taken off the develop branch, which has a bunch of jank around it. We recently switched out our camera for a cinema machine camera in order to make stuff a little bit more cinematic. But there are tons of kinks around it that we need to iron out in order to make it good.

Q: Players can currently have a maximum of 10 Seedlings. Are there plans to increase or even lower this number in the future?

Honestly, there's no reason to limit it. We have seen in some early playtests where people acquired 40-50 Seedlings. When I heard that, I was like, how the hell is that even possible? I didn't know our UI supported that.

Maybe that's the reason why we backed down to 10. I think one of the lessons we also took from previous game development experiences is that having a variable degree of engagement for different players is a good thing. If you look at most MMOs, you can split up the user groups into certain user types.

You might have a guild leader that spends a lot of time playing the game. You might have somebody that is dabbling in trade that is somewhere there in between. You might have somebody that comes in for a combat session every now and then that has lower engagement.

You want to create the possibility for all of these players to be able to spend as much or as little time as they want.

Q: SEED throws a lot of information at the player at once. Are there plans to change or refine how the information is given to them?

Absolutely. Our goal is to make a massively multiplayer simulation game that is accessible to the mass market. Right now, we are focusing a little bit more on the more hardcore user groups. SEED has three user personas: caretakers, entrepreneurs, and governors.

I do believe the beginning push is more towards business players and entrepreneurs. For a caretaker to come into this Tamagotchi on Stairways experience where they can see their Seedling live out a lifetime, there needs to be some scaffolding for them to be able to do that.

What we're seeing is that cities that are more established or settlements that are more established and have created certain services and started creating certain jobs, those are the things that will allow these more casual players to enter the game more smoothly. Definitely, the onboarding is now pretty much what we are focusing on until we go live. But again, we will be focusing a little bit more on not the hyper-casual player, but somewhere in-between, in order for them to create this foundation where people can come in and actually thrive.

Q: One thing I learned during the preview is that Seedlings players form dictatorships where one player can have government control over the others. What prompted the team to include this option into the game?

Well, it's a really interesting real-life problem and our game is all about real-life problems. Collective decision-making is very difficult.

It's very difficult even in a small company like a client. Collective decision-making is a struggle sometimes. It's a subject that we have been very interested in: how do people self-govern?

It's something that, I would say, was also part of the problem with EVE Online, which is that governance was really static. So I would say that you have a corporation in EVE Online and you have a CEO, and the CEO stops playing the game. Then the corporation does some things for granted. It's kind of stuck. They either have to disband and regroup or do something about it.

I think that any group that you create, it always changes over time. You can go from a small friend group to a large friend group or whatever. These dynamics are always shifting, so to speak.

You might even say that there would be no peace without war, right? There always needs to be something. There's no good without evil kind of idea. And I'm not saying that dictatorships are evil or anything like that.

But, having a broad possibility of players to be able to choose how to self-govern. Anything from having a limited democracy to a hardcore dictatorship. Providing players with those tools and just seeing what they do with them is a very fascinating question, I think.

And I honestly think that if we do things right, we might even learn something about collective decision-making. I'm not saying that we absolutely will. But it is obviously happening in a virtual world where stakes are different and whatnot.

But yeah, I truly believe if we give players these options, they will exploit them in certain ways, and we will see results that we haven't seen before. And the scary thing actually is that right now, dictatorships seem to be the way to go. Because the most successful societies in SEED right now are dictatorships.

Q: Are there concerns that the late-stage gameplay in SEED will become too complex and take away from the player simulation aspect of the game?

I think not, because each Seedling has a certain lifetime. So they are born and they live for an X amount of time. So I definitely think that that loop of caring for a life and guiding them through some lifetime, whether that be supporting a sad janitor or treating some mega doctor or whatever, I think all of those stories are fascinating and also something that people can learn about.

Q: Obviously, the intention now is to complete SEED, but what plans do you have to support SEED after its release?

This is just the very beginning.

Internally, the SEED that we are releasing is called 0.3. We are just going into the early steps of what this is. This has been something that we have pursued for many, many years now. We started talking about this, I believe, 23 years ago now.

So it's certainly been a journey and something that we really want to follow through with and keep adding to and keeping alive. We are trying to create a parallel world that lives alongside ours.