What the Data Shows

Burton's study examined 9,879 paid Steam releases after filtering out spam, free-to-play titles, and purely commercial releases . Of those, approximately 21% carried an AI disclosure — a rate that had been accelerating throughout the year .

The headline finding: AI-disclosing games averaged just 4 reviews in their first post-launch month, compared to 7 reviews for games without AI disclosures . Nearly 20% of AI-tagged games received zero reviews at all in that period . Among games that did accumulate at least 100 reviews, AI-disclosing titles had a lower median positive review percentage: 84.6% vs. 88.3% for traditional games .

Burton used a causal statistical model that controlled for confounding variables such as developer experience, publisher backing, marketing investment, and genre to isolate the effect of AI disclosure itself . The model estimated that, all else being equal, declaring AI use was associated with cutting a game's potential audience roughly in half .

The Penalty Is Worse for Established Developers

Multiple sources report that the sales penalty is more severe for established studios that disclose AI use . The same causal model suggests that experienced developers face an estimated 40–60% drop in sales, while inexperienced developers — who have lower baseline expectations from audiences — see a comparatively smaller penalty .

Burton's analysis indicates that without an AI disclosure, established studios would have enjoyed a 20% to 60% sales boost from their existing reputation . The finding aligns with the broader argument that consumers tend to interpret heavy reliance on AI as a signal of low design effort or corner-cutting .

Broader Context: Valve's Evolving AI Disclosure Policy

Valve first required AI disclosure on Steam in January 2024, introducing two categories: pre-generated AI content and live-generated content . The policy was added to the Steam Distribution Agreement and required developers to state that their AI-powered games did not include illegal or infringing content .

In January 2026, Valve "significantly rewrote" the policy to narrow the scope of disclosure obligations . Under the updated rules:

• AI-powered development tools used behind the scenes, such as code assistants, concept ideation tools, or workflow efficiency software, generally do not require disclosure .
• AI-generated content consumed by players — including artwork, audio, localization, narrative elements, marketing assets, and content visible on a game's Steam page — must still be declared .

Valve's updated submission form explicitly states: "We are aware that many modern game development environments have AI powered tools built into them… Efficiency gains through the use of these tools is not the focus of this section" .

The policy revision was widely seen as a response to developer backlash and the recognition that blanket AI labels were harming commercially viable games . By January 2026, approximately 8,000 Steam titles had disclosed generative AI use in the first half of 2025 alone, compared to roughly 1,000 in all of 2024 .

Industry Sentiment and the 'AI Stigma'

The Game Oracle study's findings have reinforced a growing phenomenon analysts call the "AI stigma" among Steam players . An academic study published in May 2026 found that players did not reliably distinguish between AI- and human-designed game levels, yet their beliefs about the creator were strongly linked with their reported experiences — suggesting that the label itself, not the actual output quality, drives negative reception .

Industry figures have weighed in. Epic Games CEO Tim Sweeney criticized Valve's disclosure mandate as "becoming increasingly irrelevant," arguing it "makes no sense for game stores" where a growing number of titles now use AI tools . The data suggests that players are penalizing AI-tagged titles regardless of actual game quality, creating a dilemma for developers who want to use AI tools without alienating their audience .

What This Means for Developers

Burton's study and the subsequent policy changes present a complex picture for developers:

1. Avoiding disclosure is not a reliable strategy. Valve's 2026 revision clarified that AI-generated content visible to players must still be disclosed. Hiding AI use carries its own risks if discovered.

2. The stigma is real and measurable. Games that disclose AI use lost roughly half their potential review volume after controlling for other factors .

3. Existing reputation works against established studios. High-budget or well-known developers face the steepest penalties, as audiences hold them to higher standards .

4. The policy landscape is shifting. Valve will continue to refine its disclosure requirements, and platforms like Epic Games Store may take different approaches .

As generative AI becomes embedded in standard development workflows — code assistants, asset generation, localization — the tension between transparency and commercial viability is likely to intensify. The Game Oracle study provides the clearest evidence yet that, for now, Steam players vote with their wallets against AI-labeled games, regardless of the actual quality of the final product.

What the Data Shows

Burton's study examined 9,879 paid Steam releases after filtering out spam, free-to-play titles, and purely commercial releases . Of those, approximately 21% carried an AI disclosure — a rate that had been accelerating throughout the year .

The headline finding: AI-disclosing games averaged just 4 reviews in their first post-launch month, compared to 7 reviews for games without AI disclosures . Nearly 20% of AI-tagged games received zero reviews at all in that period . Among games that did accumulate at least 100 reviews, AI-disclosing titles had a lower median positive review percentage: 84.6% vs. 88.3% for traditional games .

Burton used a causal statistical model that controlled for confounding variables such as developer experience, publisher backing, marketing investment, and genre to isolate the effect of AI disclosure itself . The model estimated that, all else being equal, declaring AI use was associated with cutting a game's potential audience roughly in half .

The Penalty Is Worse for Established Developers

Multiple sources report that the sales penalty is more severe for established studios that disclose AI use . The same causal model suggests that experienced developers face an estimated 40–60% drop in sales, while inexperienced developers — who have lower baseline expectations from audiences — see a comparatively smaller penalty .

Burton's analysis indicates that without an AI disclosure, established studios would have enjoyed a 20% to 60% sales boost from their existing reputation . The finding aligns with the broader argument that consumers tend to interpret heavy reliance on AI as a signal of low design effort or corner-cutting .

Broader Context: Valve's Evolving AI Disclosure Policy

Valve first required AI disclosure on Steam in January 2024, introducing two categories: pre-generated AI content and live-generated content . The policy was added to the Steam Distribution Agreement and required developers to state that their AI-powered games did not include illegal or infringing content .

In January 2026, Valve "significantly rewrote" the policy to narrow the scope of disclosure obligations . Under the updated rules:

• AI-powered development tools used behind the scenes, such as code assistants, concept ideation tools, or workflow efficiency software, generally do not require disclosure .
• AI-generated content consumed by players — including artwork, audio, localization, narrative elements, marketing assets, and content visible on a game's Steam page — must still be declared .

Valve's updated submission form explicitly states: "We are aware that many modern game development environments have AI powered tools built into them… Efficiency gains through the use of these tools is not the focus of this section" .

The policy revision was widely seen as a response to developer backlash and the recognition that blanket AI labels were harming commercially viable games . By January 2026, approximately 8,000 Steam titles had disclosed generative AI use in the first half of 2025 alone, compared to roughly 1,000 in all of 2024 .

Industry Sentiment and the 'AI Stigma'

The Game Oracle study's findings have reinforced a growing phenomenon analysts call the "AI stigma" among Steam players . An academic study published in May 2026 found that players did not reliably distinguish between AI- and human-designed game levels, yet their beliefs about the creator were strongly linked with their reported experiences — suggesting that the label itself, not the actual output quality, drives negative reception .

Industry figures have weighed in. Epic Games CEO Tim Sweeney criticized Valve's disclosure mandate as "becoming increasingly irrelevant," arguing it "makes no sense for game stores" where a growing number of titles now use AI tools . The data suggests that players are penalizing AI-tagged titles regardless of actual game quality, creating a dilemma for developers who want to use AI tools without alienating their audience .

What This Means for Developers

Burton's study and the subsequent policy changes present a complex picture for developers:

1. Avoiding disclosure is not a reliable strategy. Valve's 2026 revision clarified that AI-generated content visible to players must still be disclosed. Hiding AI use carries its own risks if discovered.

2. The stigma is real and measurable. Games that disclose AI use lost roughly half their potential review volume after controlling for other factors .

3. Existing reputation works against established studios. High-budget or well-known developers face the steepest penalties, as audiences hold them to higher standards .

4. The policy landscape is shifting. Valve will continue to refine its disclosure requirements, and platforms like Epic Games Store may take different approaches .

As generative AI becomes embedded in standard development workflows — code assistants, asset generation, localization — the tension between transparency and commercial viability is likely to intensify. The Game Oracle study provides the clearest evidence yet that, for now, Steam players vote with their wallets against AI-labeled games, regardless of the actual quality of the final product.