David Gaider warns AI will make gamedev “frustrating as hell” and erase entry-level tasks

David Gaider does not sound impressed with the pitch that generative AI will “save time” by removing tedious busy work from game production. The former BioWare writer and lead writer on the first three Dragon Age games told GamesRadar that AI’s inconsistency makes it hard to appraise, troubleshoot, and clean up what it produces, because teams often have to “go back and touch up its output,” without understanding why it generated a certain result. His blunt verdict: it would be “frustrating as hell,” and “it’s not ready for prime time.”

He went further, calling out the workforce logic behind the sales pitch. Gaider said the idea that AI can replace rote tasks often handed to junior developers is “not necessarily a good thing.” His specific warning lands where boards and executives should pay attention: “How are we going to train up the next generation of devs if we eliminate every entry-level task?” In other words, even if AI helps ship some content faster, it may quietly hollow out the pipeline that teaches people how to build and debug games in the first place.

That tension shows up in the way teams adopt AI today. Gaider’s main technical concern is not that AI can generate content. It is that it is not consistent. In practice, inconsistent output forces humans into a cleanup loop: evaluate what came out, fix what is wrong, and try to infer why it failed. That is labor too, and in many production environments it happens late enough to become expensive. The second-order effect is operational: once you cannot reliably predict AI behavior, you spend more time validating than iterating, which turns “automation” into rework.

There is also a strategic framing problem. AI is commonly positioned as a way to carve away busy work, and sometimes in a “creative” way, too. Gaider acknowledged that the usual justification for early use is placeholders, or helping with early prototypes and concepts. But he is wary of even that narrower use, because artists have not agreed to have their data “pillaged.” That matters for studios beyond ethics headlines. Consent and trust influence hiring, retention, and internal morale. When creatives believe their work is being mined without agreement, the cost shows up later as friction, not just in public backlash.

Player sentiment is already acting like a market signal here. The article points to the reaction to the new Crazy Taxi game as evidence that many players are skeptical of AI’s creative applications. That kind of skepticism is dangerous because it can compress a studio’s margin for error. Games are expensive, timelines are unforgiving, and marketing spend amplifies risk. If players associate AI-used assets with low authenticity, studios may find themselves defending creative choices instead of building excitement for gameplay.

Other developers quoted in the GamesRadar feature do not all share the exact same concerns, but their themes rhyme. For instance, Iron Lung and Dusk creator David Szymanski said he is not “not categorically against AI as a whole technology,” but he thinks it goes too far to “hand wave all the ethical concerns about plagiarism, environmental impact, and job security.” That list is important because it connects “AI use” to multiple stakeholder expectations: creators want credit and consent, communities want environmental accountability, and workers want stability.

And it is not only creators speaking in abstractions. Danny Koo, Marvel Rivals executive producer, said plagiarism worries were “of particular concern,” and that the team avoided AI art tools to ensure the game’s assets weren't “poisoned.” That phrasing points to a reputational and legal fear that boards should understand even when they are not thinking about art pipelines day to day. In a world where IP claims and provenance matter more each year, “poisoned assets” can translate into production downtime, rework, and potential takedowns, not to mention a damaged brand promise to players and partner studios.

Put all of this together, and Gaider’s warning stops being only about writers and junior developers. It becomes a broader capability question for any studio adopting AI: are you replacing tasks, or are you replacing skills? If AI output forces constant human correction, you may not reduce labor at all. If AI reduces entry-level opportunities, you risk shrinking the next generation of developers who understand the messy reality of game systems, content pipelines, and debugging. For executives, the strategic stake is straightforward: the quickest way to cut costs today can be the fastest way to lose velocity tomorrow, because you weaken both the production process and the talent pipeline that makes games work in the first place.