Mobile Mavens: What Google’s AI world-building tech Project Genie means for games
- “If successful, it will help powering more games faster and saves time for game developers." - Pascal Clarysse
- “People should not see it for more than what it is - an early experiment - and Google themselves have made this clear in their gated rollout.” - Rebecca Liao
- “For traditional game developers, this is a moment to actively engage rather than resist.” - Wenfeng Yang
Google's AI-driven Project Genie has been a hot topic among games industry professionals, thanks to its ability to generate short, interactive virtual environments from text prompts and images. Its arrival triggered share price drops across several major game companies, with reactions ranging from enthusiasm to scepticism.
To find out more about the industry thinks, we asked our Mobile Mavens to discuss their thoughts on the tech and what it means for games development and the wider market.
Pascal Clarysse
Google Genie is yet another neat little tool we can add to our arsenal of AI-based prototypers and A/B testing at ideation and incubation stage of game development. Like Figma and Lovable, it can find its place in the middleware pipeline that many game developers will use at an early stage of a new project, when it's time to quickly test new ideas and benchmark them against one another.
Through the improvements of Agentic AI, Google Genie can graduate into something more meaningful that actually saves time to developers, designers and artists at scale in their day-to-day output, like Layer.ai for visual assets, ElevenLabs for voice cloning and localisation, or Claude as a coding assistant.
If successful, it will help powering more games faster and saves time for game developers. I can't quite imagine a scenario in which it represents a threat to Take-Two, Roblox or Unity though.
“If successful, it will help powering more games faster and saves time for game developers.”Pascal Clarysser.
The share price drops of gaming companies on the announcement day of Google Genie is difficult to understand. Unity will remain one of the main engines of choice for game developers, wherever they vibed the prototypes before getting into actually building the commercial game, and regardless of how the procedural environment is generated.
Take-Two's GTA 6 is certainly nowhere near under threat by this, and is the most puzzling dip. Surely, strong IPs and quadruple-A level titles are precisely the type of assets for which the fused-zeitgeist future is shining bright.
Maybe Roblox, as the youth's social platform of choice might lose a dent to this "other leisure"? In the entertainment space, we are all ultimately competing for people's attention after all. I can hear this argument in theory, but it only materialises in practice if (big if) Google builds more layers and turns Genie into a social gaming platform capable of competing, then yeah this could theoretically into another time-consuming emerging medium.
But this is nowhere near what Genie is right now, and not even what Google announced it to be - their positioning is firmly aimed at Agentic AI use cases. If anything, this only reinforces the mega-trend that the future of entertainment is "creators-first, not just consumers" and this only comforts Roblox in their lead as the best positioned company for that version of the future.
“The new Google gadget is definitely not worth dropping game companies' shares in panic.”Pascal Clarysser.
Conclusion: the new Google gadget is definitely not worth dropping games companies' shares in panic. This reaction shows once again the wide disconnect between stock exchange investors and actual operators on the field
Be of note that this kind of potential "emerging medium" is not born with Google Genie 3 and there is a cool startup from Helsinki that is currently ahead of the curve compared to Google's release.
Check out StarStuff. There you can generate experiences in open worlds, with consistent avatars and more advanced gameplay, alongside a community of one million users and 72K creators - shoutout to Hanna Toivonen and her team. They're doing great work. Yet despite their strong start, I'm sure even they won't take shade from me saying that GTA 6 and Roblox still remain hotter properties under the sun for the foreseeable future.
Martine Spaans
As a first reaction I am both mind-blown and nervous at the same time. What would it mean if creating a game is really this easy? Probably an even bigger overload of mediocre games without a proper storyline or refreshing and balanced gameplay. With every new AI-creation tool released it becomes clear over and over again that people are interested in things made by people, whether it be music, art, stories, movies or games.
“With every new AI-creation tool released it becomes clear over and over again that people are interested in things made by people.”Martine Spanns.
And of course there are teething problems. A first wave of reactions was already released over the weekend around the legal issues this tool would cause if it would go live this way, but surely a company like Google would be able to tackle that issue if they have any interest in that.
It’s only available to US AI Ultra users so far, so no way to try it out yet, but I am certainly curious if it’s as easy as the marketing video promises.
Wenfeng Yang
In the short-term, Genie 3 will have a real impact on the game industry, especially on game engine companies. But from an investor’s perspective (yes, i am an investor now!), it also opens up a large number of new paradigm opportunities.
“From an investor’s perspective It also opens up a large number of new paradigm opportunities.”Wenfeng Yang.
On top of Google Genie 3, we will likely see many application-layer companies emerge. This is not limited to games - it also creates new windows of opportunity for video, short-form drama, interactive films and game-like narrative experiences.
For traditional game developers, this is a moment to actively engage rather than resist. By injecting their data and experience into these new tools and continuously training the models, they can enable creators who are not originally from the game industry to produce higher-quality works with stronger personal expression.
Rebecca Liao
Genie 3 is a truly impressive world-building prototype engine that will open the doors to more generative AI experiments. Its ability to be responsive to even casual prompting is better than other similar prototype products in the market. People should not see it for more than what it is - an early experiment - and Google themselves have made this clear in their gated rollout.
“If game developers are looking for a Unity or Unreal replacement, this is definitely nowhere near that.”Rebecca Liao.
If game developers are looking for a Unity or Unreal replacement, this is definitely nowhere near that. In addition, the worlds that a user builds using Genie are not exportable as code, so they cannot be used to save production time on a game.
What Genie can eventually evolve into is another short-form video content generator, where its worlds can be spliced with other content for a richer experience. In addition, Genie can eventually create hyper-casual games with beautiful graphics, but the game loop will need to be incredibly sharp to grab a player.
One thing that Genie does highlight is the need for more efficient storage and recall on memory in AI training models, something that is a top technology challenge across AI, not just gaming and entertainment. The reason Genie is slow to respond to input is that it is processing a huge amount of information in order to get the AI to accurately reflect the move and adjust the surrounding world correctly.
Everyone knows lag is the killer in gaming, so if Genie does hope to be a tool in this space instead of non-interactive entertainment, that should probably be the first problem to solve for the next iteration.
Will Luton
My general feeling about these kinds of prompt-to-game services is that they’re going to be most impactful for UGC. I can see a new platform in the style of whatever we call Fornite and Roblox now metaverse is ick-inducing, where the barrier to creating content is reduced down to prompts or voice commands. And from that we’re going to see some cool social experiences.
“The reality of anything AI is that it all falls down when you want specificity or complexity.”Will Luton.
The reality of anything AI is that it all falls down when you want specificity or complexity. And games are full of very complex and very specific interactions. Game feel is so important and requires a lot of dialling in and you don’t get it if there’s too much abstraction.
Using AI tools to make games right now is a lot like trying to play the violin while wearing padded mittens. So, unless there’s a massive change in the way generative AI works where there is persistence and specificity in iterations, these kinds of tools will remain the focus of hobbyists.
Jacki Vause
Project Genie, like Sora, is overhyped and looks very pretty. It’s (as it is titled) a project, it’s not a game and it can’t be monetised, it's just showing a leading edge AI approach to world creation and interaction.
When Sora came out everyone was going mad for it but, for many months, we just saw demos and those demos were all that was available for ages. A lot of people are riled because it caused a wipe out on tech stock prices.
But those of us who have ridden this beast for a while see that happen a lot when the markets get a bit of a kick and re-examine old assumptions. Others are asserting (rightly) that Google’s track record when it comes to games projects isn’t stellar (Stadia, Arcade, Instant Games etc.).
“When Sora came out everyone was going mad for it but, for many months, we just saw demos and those demos were all that was available for ages.”Jacki Vause.
However, Google won’t just be looking at AI worlds from a games lens and what Genie does tell us is that AI is here to stay and we have to accept that and now get ahead of everything that comes with it and that can be fundamentally affected as a result of it - ethics, copyright, investment, people - the list is too vast to go into one comment.
I anticipate that the games industry will continue to innovate and show a lead for mobile apps and Project Genie will be part of a number of tools and innovations like Ludo.ai's suite of tools for game devs or Scenario's creative workflow tools that are already being used on a daily basis our rich ecosystem.
Oscar Clark
For me, this kind of thing is super interesting but I reserve judgment too! Making an expansive environment which feels in character with the world you want to create is immensely interesting, and I can see tools like this could speed up the process tremendously.
“A lot of what makes games work goes beyond this simple 'state' experience.”
Oscar Clark.
However, as seductive as this could be, it's not the only characteristic needed to make a game. Players need to learn an environment, which means once we have defined a space, we need to use our navigational memory to return. Gen AI notoriously doesn't really work that way.
Additionally, a lot of what makes games work goes beyond this simple 'state' experience. Locations are about what we do in them, and interacting in the world matters as much as the intentionality of the environmental design.
A massive lesson for me from working on PlayStation Home was the role of the environment as part of UX. Environment artists bring this understanding into a game's playability; they communicate story and context and indicate play direction, even down to creating pathways that instinctively guide the player to goals/actions. For example, when a player is running from a horde of zombies and turns left, not realising till later that there never was a right turn.
“Super interesting, but there is a lot more I'd need to see to have confidence that this will immediately change how we make games.”
Oscar Clark.
Add to that this is not a game engine native; needed to support the other layers of the game. That matters as if we are to maintain an experience we need to have some control over the tools - especially for live ops. There are some tools trying to do this within the game context such as Tilki however.
In summary - super interesting, but there is a lot more I'd need to see to have confidence that this will immediately change how we make games. That being said I think tools like this will deliver some really interesting innovations in how quickly we can realise new gaming concepts.
Stuart De Ville
From an indie developer’s perspective, Genie 3 is a compelling technical achievement, but it raises important questions about the nature of creativity in games production.
World models of this kind are highly effective at generating visually and spatially coherent environments, yet coherence alone is not what distinguishes memorable game art.
“World models of this kind are highly effective at generating visually and spatially coherent environments, yet coherence alone is not what distinguishes memorable game art.”Stuart De Ville.
Distinctive, resonant work emerges from human authorship, taste, and a deep understanding of craft, qualities that are developed through practice rather than inferred statistically.
It is also important to contextualise tools like Genie 3 within the broader history of independent development. Indies have always relied on shortcuts and enabling technologies, from asset stores and middleware to auto-rigging, animation libraries, and procedural systems.
These tools have long been accepted as pragmatic responses to limited time, budget, and team size. In this sense, so-called ‘AI’ is not a rupture, but an extension of an existing toolchain.
There is no artificial intelligence acting with intent or authorship here, only increasingly sophisticated algorithms operating within human-defined constraints.
Where Genie 3 and similar systems show genuine promise is in rapid prototyping and exploratory design. A significant proportion of game development consists of work that is ultimately discarded, and accelerating this phase without consuming the time and energy of skilled artists represents a meaningful productivity gain.
“Used appropriately, these tools can help teams validate ideas, test spatial concepts, and explore mood or scale before committing to costly production pipelines.”Stuart De Ville.
Used appropriately, these tools can help teams validate ideas, test spatial concepts, and explore mood or scale before committing to costly production pipelines.
The risk lies in misunderstanding both the tool and the landscape in which it is deployed. When speed is mistaken for creativity, or when generative outputs are treated as authored work rather than provisional material, the result is overuse and creative flattening.
If AI-generated content moves from exploration into final production without human intention and refinement, we risk a wave of visually competent but culturally indistinct games. In an increasingly crowded market, intentional, hand-crafted work remains a key differentiator, not as a rejection of new tools, but as a disciplined and informed use of them.
Companies
People