CD Projekt Co-CEO Questions Fully AI-Generated Games

Michal Nowakowski, joint CEO of CD Projekt, told Edge's Knowledge newsletter that he "knows for a fact" games made entirely with generative AI are coming, recounting conversations with founders of AI-first studios who described rapid prototyping and very fast launches (reported by Yahoo/Tech and Edge via Yahoo). Nowakowski also said he "has some doubts whether this is really the path to follow," using that phrasing in the Edge interview as quoted in coverage by GamesRadar and Yahoo/Tech. Rock Paper Shotgun reports that in the same conversation Nowakowski referenced Epic Games integrating models such as Claude and Gemini into Unreal Engine 6, and said CD Projekt has had unusual access to tinker with the engine's internals, per RPS reporting. RPS and GamesRadar additionally relayed concerns voiced by publishers and external-development leads at Digital Dragons, including a quote from 11 bit external development director Rufus Kubica about how heavy AI use in prototypes raises questions about whether teams actually have the artists to realise a promised quality.

Industry observers note that the fastest generative-AI-driven prototyping workflows trade human time for automated content generation, allowing large numbers of design variants to be produced quickly. This increases the volume of candidate concepts hitting early testing, but it does not automatically resolve downstream integration tasks such as animation pipelines, bespoke narrative design, QA, and optimisation for platforms. Observers also point to the distinction between AI-assisted authoring workflows and attempts at end-to-end, fully generated products; the former is already being integrated into engines and tooling, while the latter presents harder technical and production challenges.

Reporting by RPS and GamesRadar highlights recurring industry concerns: rights and licensing of training data, the ergonomics of prompt-driven content creation, and publisher risk assessment when prototypes are AI-generated. Multiple outlets quoted industry figures asking how a studio can demonstrate the capacity to deliver on a prototype if the visible deliverable was largely produced by generative models. Those questions map to hiring and vendor-evaluation processes, not to any specific claim about CD Projekt's internal plans.

Companies adopting generative tools for early-stage design typically gain iteration speed, but face friction at handoff points where AI output must be polished into a production-quality game. Legal and ethical issues around scraped training data and attribution remain unsettled, and publishers reported in coverage are already factoring those risks into commissioning decisions. The attention economy angle Nowakowski mentions, as reported by Yahoo/Tech, compounds the commercial risk: a flood of rapid, low-cost titles increases competition for player time and discoverability.

• •Track announcements from engine vendors on how models will be integrated into authoring tools and what access partners receive, following RPS reporting about Unreal Engine 6 integrations.
• •Watch publisher procurement language and external-development RFPs for new clauses addressing AI provenance and demonstrable artist capacity.
• •Monitor early commercial releases from studios claiming heavy generative-AI pipelines to see whether post-launch support and quality match studio promises.

Per the reported interview, a senior CD Projekt executive expects a surge of AI-first games but expressed explicit doubts about whether end-to-end generative production is commercially or creatively preferable. Industry observers and publisher representatives quoted in the coverage frame delivery capacity, IP, and discoverability as the main operational concerns to watch as the technology diffuses.