European game worker unions led by IWGB have published a manifesto against generative AI in game development, calling on workers, studios, regulators, and players to resist its spread.

Image: gamereactor.eu
European game unions put players inside the AI fight
Six European game worker unions have published a joint manifesto opposing the spread of generative AI in video game development, with one of its sharpest demands aimed outside the workplace: players should “support games made without genAI” and be vocal when they dislike AI in games.
GamesIndustry.biz reports that the manifesto, titled “On Generative AI Usage in the Games Industry,” is led by Great Britain’s IWGB Game Workers Union in collaboration with members of Game Workers Unite Ireland in the Republic of Ireland, Spain’s CSVI, France’s STJV, Germany’s GDRT, and Italy’s FIOM-CGIL Milan Work Council. Rock Paper Shotgun and Kotaku also report that the document was published around a European game worker union summit in Manchester.
The concrete demand is simple, but the strategic move is bigger. Until now, much of the public fight over generative AI games industry adoption has played out as an internal production question: whether AI tools save time, whether they replace artists, writers, coders, QA staff, translators, or junior workers, and whether studios can use them without legal or ethical blowback. The IWGB-led manifesto reframes that as a market question. If players begin treating “made without genAI” as a buying signal, publishers do not only face labor resistance. They face a consumer trust problem.
What the manifesto says generative AI is doing to game work
The manifesto’s central claim is direct: “GenAI is currently harming the games industry, its workers and the games they produce.” The unions argue that, left unchecked, this harm will reach more workers, more games, and more players. The document defines genAI as systems that generate text, pictures, video, or music from limited inputs, usually prompts, using LLMs and other energy-intensive models.
The manifesto is focused on game development rather than the whole AI economy, but it explicitly points to wider harms including severe environmental damage, destruction of communities, and exploitation of the Global South. Within games, its scope covers both assets players may see in a finished release, such as art, sound, code, and dialogue, and materials used during production, including concept art, references, and pre-translation work.
The unions list five broad areas of concern, as summarized by GamesIndustry.biz: worsening working conditions, damaged worker-employer relations, appropriation of workers’ output, discrimination, and hidden costs. The document says genAI is directly and indirectly replacing workers, that some game workers are being forced to use it explicitly or implicitly, and that using it can reduce the sense of ownership workers feel over what they produce.
That last point matters because games are collaborative systems built from hundreds of judgment calls. If a tool changes who gets to make those calls, who gets credit for them, and who learns from them, it changes the studio’s talent pipeline. The manifesto argues that genAI is contributing to “a marked decline in junior positions within the industry,” creating a risk that studios optimize for short-term output while weakening the next generation of experienced developers.
The worker asks: refuse, document, organize
For workers, the manifesto’s recommendations are deliberately practical rather than abstract. Rock Paper Shotgun reports that it tells workers to avoid genAI services whenever possible, distance themselves from genAI usage, and make clear to employers why they do not want to use it. The manifesto also urges workers to organize and unionize before they are already under threat.
The document’s logic is that “choice” is often weaker in practice than it sounds in policy. The manifesto says some companies may not formally require genAI use, but can set targets that workers cannot meet without it. In that environment, a tool advertised as optional can become functionally mandatory. From a production-strategy perspective, that is the same pattern seen in other workplace systems: once management tunes schedules around a tool’s assumed productivity gain, refusal becomes a performance risk.
The unions also argue that genAI changes the texture of labor itself. Rock Paper Shotgun notes the manifesto’s claims around worker fatigue, cognitive decline, burnout, and loss of ownership. The document says work involving genAI can be less desirable, fulfilling, and educational than doing the same work without it. That is a particularly important claim for entry-level roles, where the job is often as much about learning the craft as delivering a task.
The manifesto does leave room for worker agency in principle. Kotaku reports that it states workers should be the ones deciding when, where, and how to use generative AI, while also weighing environmental, cost, ethical, legal, honesty, and artistic concerns. The tension is clear: the unions are not presenting genAI as a neutral productivity patch that can be safely rolled out from the top down. They are treating it as a power shift that workers must be able to contest.
The studio asks: do not make genAI a hidden production default
For employers, the manifesto’s recommendations push in the opposite direction of the current software sales pitch. Rock Paper Shotgun reports that it urges studios to ensure no worker is required to use genAI and to disallow genAI in company codes of conduct, including during early development. GamesIndustry.biz adds that the recommendations include training staff to identify genAI use and involving employees in AI policy decisions.
The early-stage point is easy to underestimate. Many publishers and managers are likelier to defend genAI as harmless when it is used for mood boards, reference material, pre-translation, rough code assistance, or pitch decks rather than final assets. The manifesto rejects that distinction as a safe harbor. Its scope explicitly includes materials that players will never see but that shape production.
That has direct implications for disclosure and compliance. If a studio says its shipped game contains no AI-generated assets, but its concepts, prototypes, or internal translations leaned on genAI, the manifesto’s framework would still treat that as part of the production problem. This is where publisher policy becomes tricky. A clean marketing claim requires clean internal tracking, and the further genAI spreads through everyday tools, the harder that tracking becomes.
The manifesto also argues that genAI can damage worker-employer relations because it is sold to employers as an alternative to workers. Rock Paper Shotgun reports the unions’ concern that AI systems’ sycophancy and lack of real-world context can encourage unreasonable requests when used to set policy, timescales, and workloads. In plain production terms, the fear is that management starts balancing the schedule around imagined machine capacity rather than actual team capacity.
The player pledge turns taste into leverage
The most consumer-facing line in the manifesto is its call for players to support games made without genAI and to be vocal about disliking genAI in games. Kotaku highlights that phrase in its coverage, and it is the part most likely to travel beyond union circles because it gives players a simple action: reward human-made work and make AI use reputationally costly.
That does not mean consumers can currently verify every claim. The sources provided do not point to a universal labeling system for games made with or without genAI. The practical route for players, for now, is to watch studio statements, store-page disclosures where they exist, developer FAQs, award eligibility debates, and post-release reporting. The manifesto’s pledge works less like a certification program today and more like an attempt to create demand for one.
The pressure point is publisher risk management. If enough players ask whether a game used genAI, studios may have to answer before launch rather than after controversy starts. If creators, critics, awards bodies, and storefronts begin treating disclosure as normal, silence becomes its own signal. That is how a soft norm can harden into a market rule without waiting for legislation.
Rock Paper Shotgun connects the manifesto to a statement by a group of IGF jurors who said they would not vote for any game using genAI in any capacity. Kotaku also notes the timing after House House confirmed that “under 0.1 percent of Big Walk’s code” was LLM-assisted. Those examples show the debate has moved from hypothetical AAA cost-cutting into concrete questions about awards, indie credibility, and how much AI use is enough to matter.
Regulators and industry bodies are already moving, but not in one direction
The manifesto includes recommendations for regulators as well as workers, employers, and players. GamesIndustry.biz reports that it references the UTAW Tech Workers’ Inquiry 2026, which includes policy proposals concerning generative AI use. That places the document alongside a growing regulatory and contractual environment rather than outside it.
The industry is already seeing different kinds of responses. GamesIndustry.biz notes that SAG-AFTRA has placed restrictions on the use of AI with a performer’s voice or likeness. The same outlet also reports that UKIE has partnered with law firm Sheridans on an AI Compliance and Accreditation Service intended to support UK studios in using AI lawfully and responsibly.
Those approaches do not all aim at the same outcome. Compliance frameworks can make AI use safer or more legally defensible for studios. The union manifesto is more oppositional, saying workers and players do not have to accept genAI use as inevitable. GamesIndustry.biz quotes the manifesto arguing that if genAI were simply a useful tool, workers would have easier lives and more free time, but “the opposite is happening.”
That is the core conflict publishers now have to manage. One path treats genAI as an unavoidable part of the production stack that needs rules, licensing, and disclosure. The manifesto’s path treats normalization itself as the danger, especially if tool owners spread adoption before raising prices and locking studios into dependency. For an industry already under pressure from layoffs, tight margins, and volatile project scopes, that dependency argument is likely to resonate with workers who have seen efficiency promises convert into smaller teams and heavier workloads.
The unanswered question is who gets to set the new standard
The IWGB game workers union manifesto does not create a binding rule for the industry. It does create a shared position across multiple European unions, and it gives workers, studios, regulators, and players a vocabulary for resisting generative AI in video game development. Its strongest bet is that AI adoption can be slowed by combining workplace organizing with consumer refusal.
For players, the immediate guidance is limited but useful: if genAI use affects your buying decision, ask studios and publishers directly, support teams that disclose human-led production practices, and treat vague answers with caution. For developers, the manifesto’s guidance is clearer: avoid genAI where possible, make objections visible, push for worker involvement in policy, and organize collectively rather than negotiating tool mandates alone.
The publisher-side calculation is changing. GenAI may promise lower costs, faster iteration, and expanded output, but the manifesto argues those gains come with labor harm, cultural homogenization, data appropriation, environmental costs, and weakened trust. If “made without genAI” becomes a label players actively search for, it could become a competitive advantage rather than a private production preference.
That is the strategic shift to watch. AI tools are spreading through the game development meta, but metas are shaped by counters. This manifesto is an attempt to define the counterplay before genAI becomes the default build.
