The digital storefront, once a curated garden of developer creativity, is undergoing a seismic shift. As generative artificial intelligence tools become more accessible, the barrier to entry for game development has plummeted, leading to a massive influx of titles flooding the world’s largest PC gaming platform. According to a compelling new study by former Habbo Hotel developer Sulka Haro, we are currently witnessing an "adoption ramp" that could see AI-disclosed titles account for over 50% of all new releases on Steam by the year 2028.
This shift presents a complex challenge for Valve, independent developers, and consumers alike. As the storefront struggles to filter a rising tide of low-effort projects, the industry is grappling with a fundamental question: What does the future of game discovery look like in a world where "vibe coding" and generative assets make mass-producing games as trivial as sending an email?
The Data: A Rapid Rise from Zero
The transformation began in earnest in 2024, when Valve introduced mandatory AI-disclosure requirements for developers publishing on Steam. This policy change provided researchers with the first concrete dataset to track the integration of generative AI into the industry’s pipeline.
Sulka Haro’s analysis, which examined approximately 53,600 games released between July 2023 and July 2026, paints a stark picture. Before the disclosure requirement, the industry lacked a transparent metric for AI usage. However, once the feature was implemented, the data revealed a steady, aggressive climb. Currently, roughly 31% of new Steam releases openly disclose the use of generative AI in their development process.
"It’s a real adoption ramp built on top of a baseline that started at literal zero," Haro notes in his Substack analysis. By extrapolating the current trajectory, Haro predicts that if the momentum remains unchecked, the "tipping point"—where more than half of all new games hitting the store utilize some form of AI—will be reached within the next 24 to 36 months.
Chronology of an Industry Shift
To understand how we reached this point, we must look at the timeline of generative tool democratization:
- Pre-2023: Generative AI tools were primarily in the hands of researchers and high-end tech firms. Steam games with AI assets were rare, often integrated behind the scenes without public disclosure.
- Early 2024: Following industry-wide pressure regarding copyright and transparency, Valve implements the AI-disclosure field for all developers. This becomes the primary source of truth for the platform’s demographic shift.
- Mid-2024 to 2025: The "AI-First" development movement takes hold. Developers begin leveraging Large Language Models (LLMs) and image generation tools to bypass traditional asset production, leading to a surge in titles that utilize AI for everything from dialogue trees to background textures.
- 2026 and Beyond: We are now seeing the "shovelware tail" manifest. Haro’s data suggests that the "5+ games per developer" threshold is three times higher for AI-heavy developers than for traditional studios. This indicates that AI is being used not just to create, but to multiply output, leading to a saturation of low-quality titles.
The "Shovelware" Phenomenon: Quantity vs. Quality
One of the most concerning findings in Haro’s research is the disparity in publisher behavior. While the headline number—50% by 2028—is staggering, the distribution of that content is nuanced.
Approximately 80% of the 37,000 distinct publishers on Steam have yet to ship a single game incorporating AI. Of those who have adopted the technology, 89% have shipped exactly one title. This suggests that the "slop factory" phenomenon is not being driven by a few massive, malevolent corporations, but rather by an influx of individual "AI developers" who are using these tools to spam the store with multiple low-effort projects.
"AI lowers the barrier not just to make a game but to make several," Haro explains. This creates a "shovelware tail" where platforms become clogged with clones and derivative works, making it exponentially harder for players to find high-quality, human-crafted indie games.
The Threat to Game Discovery
The implications of this flood extend beyond simple store clutter. Industry analyst Mat Piscatella of Circana has recently sounded the alarm on "AI vibe coding," a practice where non-programmers use AI to quickly assemble clones of popular games.
When a game goes viral—like the climbing game Peak—it is almost immediately met with a wave of low-effort, AI-generated "rip-offs." These clones often use the same branding, similar mechanics, and AI-generated marketing assets to confuse consumers and siphon off sales. For indie developers, this is not just an annoyance; it is an existential threat. When legitimate games are "buried" under a mountain of AI-generated clones, the economic model for small studios begins to collapse.
Industry Perspectives: Why Developers Hesitate
Despite the ease of use offered by generative AI, there remains a significant portion of the development community that is actively resisting the trend. In recent surveys, many developers cited three primary reasons for their reluctance:
- Legal Uncertainty: The murky copyright status of AI-generated assets makes them a liability for long-term IP ownership.
- Lack of Artistic Cohesion: Many developers feel that AI tools struggle to maintain a consistent "soul" or art style across a complex project, resulting in a disjointed user experience.
- Ethical Concerns: The gaming community has shown a strong, vocal aversion to generative AI, often viewing it as a replacement for human talent rather than a supplement.
The pushback is not just ideological; it is financial. Games that are perceived as "AI-slop" are increasingly being review-bombed or ignored by the community, leading to a market signal that players value human effort.
Implications for the Future
The prediction of 50% AI-generated games by 2028 is not a guarantee of doom, but it is a wake-up call for platform holders. If the current trajectory holds, Valve may be forced to reconsider how it handles storefront discovery and curation.
Potential Regulatory Responses
- Enhanced Curation: Steam may need to move away from its largely automated submission process and implement more robust quality checks to prevent AI-driven spam.
- Stricter Labeling: Beyond basic disclosure, platforms might require developers to specify exactly which parts of the game are AI-generated, allowing users to filter their search results.
- Algorithmic Adjustments: To protect indie developers, discovery algorithms may need to be adjusted to prioritize human-made, long-term support projects over "one-and-done" AI releases.
Conclusion
The data provided by Sulka Haro confirms what many in the industry have felt intuitively: the barrier to entry has never been lower, but the cost of discovery has never been higher. As we hurtle toward a future where half of the games on Steam are AI-assisted, the responsibility falls on both the platforms to protect their ecosystem and the players to remain discerning.
The "sizzling garbage" of the current AI boom is a temporary growing pain of a technological shift. However, if the industry does not address the glut of low-effort, mass-produced content, it risks alienating the very audience that sustains the medium. The future of gaming may be increasingly automated, but as the current market response shows, there is no substitute for the creative human spark that defines the medium’s best work.
