Steam Players Call for an AI Content Filter as Next Fest Discovery Frustrations Grow

Steam Next Fest is designed to be one of the most useful discovery events on PC: a limited-time showcase where players can browse demos, watch developer broadcasts and build wishlists before smaller games disappear into Steam's enormous daily release schedule. But the latest edition has sparked a familiar complaint with a sharper new edge. Many players say finding promising projects has become harder amid what they perceive as a surge of low-effort submissions and games promoted with generative-AI artwork.

The response from some members of the Steam community is straightforward: if developers already disclose AI-generated material, Valve should give customers a way to filter it out. Calls for a "no-AI" option have appeared alongside broader criticism of Next Fest's recommendation flow, with players arguing that the event's scale is no longer enough of a benefit if browsing it requires sorting through pages of material they do not want to see.

A discovery event facing a discovery problem

Next Fest has always included a vast range of projects, from polished indie releases with established audiences to early prototypes, hobby projects and games that may never launch. That openness is part of its appeal. A player can stumble across an inventive puzzle game, a striking narrative adventure or a promising strategy title that would otherwise receive little marketing attention. The downside is that the event has never been tightly curated.

For years, Steam has generally relied on its storefront systems, user interest and engagement data to determine which games rise to prominence. Popular demos can climb through wishlists, downloads, streams and community discussion, while projects that fail to attract an audience tend to sink out of view. That model can work over time, but it is less satisfying during an event built around limited attention. Players want the first few hours of browsing to feel exciting, not like a cleanup task before they can reach the games that fit their tastes.

For useful background on this topic, read Joe Ziegler Says Path of Exile's Build Freedom Keeps Him Coming Back After 3,400 Hours.

This time, community criticism has focused heavily on visual assets that appear to have been generated from text prompts, as well as projects players describe as "vibe coded" or rapidly assembled using AI-assisted tools. The concern is not restricted to whether a demo itself was produced with generative AI. For many users, promotional capsules, store artwork and screenshots are part of the problem because they shape the first impression of a game and can make the storefront feel less trustworthy or less distinctive.

One widely shared request from a Steam user asked for more control over the Discovery Queue, including the ability to exclude titles containing AI-generated content and to block unwanted genres or tags such as incremental games. The argument is that Steam already has disclosure requirements around AI-generated material, so a consumer-facing preference setting could make that information useful rather than merely present on individual store pages.

Why an AI filter appeals to players

Steam's audience is not unified in how it feels about AI-assisted development. Some players may be comfortable with limited uses, such as placeholder art, localization assistance or internal production tools. Others prefer to avoid games with any generative-AI output, whether it appears in art, writing, audio, code or marketing. A filter would not settle that disagreement, but it could reduce the need for Steam to impose a single consumer preference across its entire storefront.

That is the key distinction in the current debate. Many players are not simply asking Valve to remove every game associated with AI. They are asking for control over their own discovery feed. In principle, an optional exclusion tool would allow shoppers who do not mind AI-related content to continue browsing normally, while letting more skeptical users prioritize projects that meet their preferences.

Steam already supports an extensive set of tags and discovery controls. Users can follow tags, ignore products, browse categories and receive recommendations influenced by their activity. The platform has also made changes to user-defined tagging over time, including reducing the number of tags users can add in May. Adding an "AI-generated content" label as a browsable and excludable category would require Valve to decide how closely such a label should map to its existing developer disclosures.

That may sound simple, but the details matter. "Uses AI" can mean radically different things depending on the game. A title could use AI-generated images in a small set of cosmetic assets, or it could use generation as a central part of its dialogue, design or code pipeline. A basic on-or-off filter may satisfy players seeking a strict exclusion, yet it would not communicate the scale or type of usage for people who want more nuance.

A more detailed system could let users filter categories such as generated visual art, text, voice material or content generated during play. However, each additional category creates a greater burden for developers and moderation teams, while also raising questions about definitions and enforcement. Steam would need clear standards for what must be reported, how disclosures are updated, and what happens when a game's page is inaccurate.

Disclosure only helps if people can act on it

Valve's AI disclosure rules are intended to give customers information about how generative AI is used in a product. Yet players objecting to the current Next Fest experience argue that disclosure on a store page does not solve a discovery problem. By the time someone opens dozens of pages to check details, they have already spent the time the filter was meant to save.

That is especially relevant during Next Fest, when the sheer number of demos makes rapid browsing essential. Store art is often the first and sometimes only signal that encourages a player to click. If a user sees artwork they consider generic, misleading or AI-generated, they may skip it. Repeating that process through a long queue can create the impression that the event is overwhelmed, even if many worthwhile demos are still present further down the list.

Some players have said they would rather wait a few days and let community attention sort the showcase naturally. As demos accumulate downloads, livestream viewers, wishlists and discussion, the most compelling games may become easier to identify. That is a practical short-term strategy, but it also highlights the dissatisfaction behind the calls for better filters: the first days of a festival should ideally be when discovery is at its most rewarding, not when players feel they need to avoid the event until its rankings stabilize.

Low-effort submissions are not new, but the tools are changing

It is important to separate two related issues. Steam has long hosted games that players regard as shovelware, unfinished experiments or low-effort releases. None of that began with generative AI, and a game's use of AI does not automatically determine its quality. Small teams can make excellent games with modern tools, while traditionally produced games can still be poorly designed, misleadingly marketed or rushed to market.

What has changed is the perceived speed and volume at which a project can now acquire marketable-looking material. Generative image tools can create store-facing art quickly. AI-assisted coding tools can help inexperienced creators assemble prototypes. Newer models have also fueled attention around decompilation and mashup mods, reinforcing a wider industry conversation about how automation is reshaping game production. For Steam users, the practical fear is that an open event can be flooded faster than storefront systems can distinguish meaningful work from material that only looks presentable at a glance.

The issue arrives as larger publishers and platform holders also explore AI-driven workflows. Epic recently delayed its upcoming Star Wars shooter while pursuing an "AI-first" direction, while Epic's leadership has highlighted AI-assisted work on the Epic Games Store launcher. Capcom has likewise discussed AI as part of its REX research and development program, which is intended to support continued improvements to the RE Engine used by games including Resident Evil: Requiem and Pragmata.

Those examples do not describe the same kind of AI usage alleged by critics of Next Fest. But they show why the discussion is unlikely to fade. AI is becoming a broader production, business and design question across the industry. As more companies adopt different forms of automation, players will increasingly ask for transparent information and tools that let them decide what they want to support.

What Valve could do next

An optional AI-content filter is one possible answer, but not the only one. Valve could improve discovery by making disclosure information more visible in browsing results, adding clearer labels to festival pages, or allowing people to customize queues with more granular exclusions. It could also alter the way Next Fest is presented, emphasizing personalized genre selections, curated collections or stronger quality signals without turning the event into a heavily gatekept showcase.

Any intervention would involve tradeoffs. A stricter submission standard might reduce spam, but it could also make it harder for new developers to reach an audience. A broad AI tag could empower consumers, but developers may argue that it flattens important differences between a few generated promotional assets and a game built extensively with generative systems. Leaving everything to the algorithm, meanwhile, preserves Steam's open approach but risks exhausting the people the event is meant to serve.

For now, the loudest message from frustrated players is less about demanding a universal ban than demanding better choice. Steam Next Fest remains a major opportunity for developers, including the human-led teams players say they are trying to find. Whether Valve responds with an AI filter, stronger labeling or another discovery tool, the next step will likely be judged by a simple standard: can players locate the demos they are excited about without feeling buried by the rest?