Imagine typing, “Make a racing game where a delivery bike dodges runaway shopping carts,” into your browser and getting a playable game. After a test run, you could ask the AI to sharpen the turns or move the race to another neighborhood.
Google's Playground, launched October 7, lets anyone create games using text prompts, with no coding required. Unity plans to introduce more advanced creation tools through its upcoming Spark product, which will connect to Playground. The partnership could attract new game developers and help Google compete with platforms such as Roblox.
What you can build today
Playground is available in the U.S. to adults aged 18 and older. Users can generate browser games from their own ideas or starter prompts, then adjust gameplay rules and physics through follow-up instructions. They can also change characters or redesign environments before sharing their games privately or publishing them in the Explore gallery.
The platform is free to use, although weekly token allowances limit how much users can create and revise. Eligible paid subscriptions increase those allowances, according to Google.
Meanwhile, Unity Spark is expected to support more complex gameplay and higher-quality 3D graphics through Unity's game technology. Games created with Spark will be shareable through Playground, but developers will have to wait to test these capabilities. Unity plans to launch a closed beta later in 2026.
Google has also been developing Genie, a separate AI research project that generates interactive environments as users navigate them. As explored in The Neuron's coverage of AI-generated worlds, Genie focuses on generating environments in response to user actions, while Playground provides tools for creating and publishing browser games.
Creating a game is easy. Refining it takes work.
In WIRED's hands-on test, Boone Ashworth created three playable games and found Playground's interface approachable. However, he also noticed recurring menus and music across the gallery, alongside similar combat choices. His weekly token allowance ran out before he could finish revising his creations.
Although the review represents one person's experience, it illustrates two challenges for AI game creators. Generated games may share recognizable elements, and improving them can require multiple rounds of prompts. Weekly usage limits could interrupt that process before developers resolve problems discovered during testing.
A similar revision process appeared in The Neuron's live Gemini demonstration, where a generated shooter improved after follow-up instructions addressed problems with its environment. The example illustrates why creators need to identify what went wrong and describe the changes they want.
Unity CEO Matt Bromberg makes a related argument in his explanation of Spark: “This isn’t about ‘one-shotting’.” He argues that creators still need original ideas they want to develop and polish, rather than simply recreating existing games through prompts.
For independent developers, AI could reduce the time required to test gameplay ideas. A small studio might compare two puzzle mechanics before committing resources to production, while a hobbyist could experiment with concepts that previously required programming knowledge.
Even when AI produces a functioning puzzle, developers must determine whether its challenges feel fair and whether rewards arrive at the right moments. Those decisions depend on playtesting and an understanding of what keeps people interested.
The growing number of amateur creators could also create opportunities for professional artists. Bromberg says Spark will connect to the Unity Asset Store, where creators can access artist-made assets for their games. More developers could mean additional customers for those assets, although Unity has not established how the new tools will affect artists' earnings.
Google and Unity also need to attract players
Google and Unity face competition from companies that already provide ways to create games and reach audiences. Roblox's Build, for example, lets users to turn prompts into games through its mobile app and move their projects into Roblox Studio for further development. Roblox also uses player retention in its game discovery systems.
Meta's Horizon Create and Horizon Studio are gradually becoming available through early access. Meta says published games can become eligible for distribution across its social platforms, including Facebook and Instagram. Such distribution could introduce creators' games to people who have never visited a dedicated game-development platform.
Google is approaching the challenge through Playground's Explore gallery, where player ratings and activity influence discovery. According to Unity, Google also plans to promote selected creations across its network, potentially giving developers another way to reach audiences.
These distribution features could become increasingly important as AI reduces the time required to produce games. If thousands of creators can generate playable experiences through prompts, more titles will compete for the same players. Developers may favor platforms that attract enough people to test their games, gather feedback, and build an audience.
Roblox already has an established community of creators and players, while Meta can draw on its existing social platforms. Google and Unity will need to demonstrate their creation tools can support games people enjoy beyond an initial experiment. Their discovery features will also need to help new creators reach players among a growing catalog of AI-generated content.
For people who have always wanted to make a game, Playground could provide an accessible starting point. The longer-term challenge for Google and Unity is whether those first experiments can develop into games that attract returning players, and whether creators will choose their platforms to publish them.
