CD Projekt has confirmed it uses artificial intelligence as a development tool but will not rely on AI to build complete games, drawing a clear line on automation during its first half 2026 earnings call. The company, which currently has 1,045 developers spread across multiple projects, said its teams will remain predominantly human-led even as AI becomes more useful in specific development areas.
"AI is definitely becoming a thing in game development as well; we are still predominantly using people, and this is our plan," the company said. It described AI as a tool like any other that can assist developers in certain places and could become more helpful over time.
The stance matters for an industry grappling with how deeply to integrate generative AI into creative pipelines. CD Projekt's position - treating AI as an assistant rather than a replacement - reflects a broader calculation among studios developing complex, narrative-heavy titles where maintaining creative control is commercially and reputationally critical.
No AI-generated games, especially for complex titles
CD Projekt was direct about its unwillingness to depend on AI for producing finished games. The company questioned whether such an approach would even be appropriate given the complexity of its projects. "We're not planning to rely on AI making complete games, and to be honest, especially with very complex games like ours, I don't even think this is the right or ethical path to follow," the company said.
This philosophy extends across a development pipeline that includes The Witcher 4, which targets a 2028 release and has the largest team at 519 developers. Cyberpunk 2 follows with 184 developers, a team that has grown over the past three months to support its current development phase. For IT and development professionals watching how AI for IT & Development is reshaping software workflows, CD Projekt's approach offers a concrete case study in setting boundaries around automation while still adopting the technology where it makes sense.
Earlier-stage projects and multiplayer experimentation
Project Hadar remains in the concept phase with more than 30 developers currently prototyping gameplay elements. CD Projekt expects to grow that team to around 50 by the end of 2026. "Hadar is still in the concept phase. We currently have over 30 people and expect to hire more, up to 50 developers by the end of this year," the company said.
Project Sirius, CD Projekt's multiplayer title, is also in pre-production. The company has adjusted team size to match current requirements and said the focus remains on testing different solutions. "This project is very different from our other projects because it's focused on multiplayer. So there's a lot of learning and testing in this particular one," the company said. The multiplayer focus introduces distinct technical challenges, which makes the experimentation phase especially important.
Two unannounced projects expand the pipeline
CD Projekt confirmed two unannounced projects are in active development, separate from titles still in the research stage. The Witcher 3 Wild Hunt Remastered does not count among them - the company treats it as a continuation of existing work on The Witcher 3. The Songs of the Past expansion, planned for 2027, and the Nintendo Switch 2 version of The Witcher 3 were separately capitalised as development projects.
The full pipeline now spans The Witcher 4 targeting 2028, Songs of the Past in 2027, Cyberpunk 2 in development, plus Project Sirius, Project Hadar, and the two unannounced projects progressing through various stages. For software developers evaluating how to integrate AI into their own workflows, this multi-project structure shows how a large studio allocates human talent across titles at different maturity levels while keeping AI in a supporting role. The AI Learning Path for Software Developers addresses exactly this kind of practical integration question.
Why this matters for IT and development professionals
CD Projekt's earnings call provides a real-world example of a large technical organization setting explicit guardrails around AI adoption. The company is not rejecting the technology - it is deploying it as a developer tool in specific areas - but it has made a deliberate decision to keep human developers at the center of complex software creation. For development teams evaluating their own AI strategy, the takeaway is not that AI should be avoided, but that its role needs to be scoped against the complexity of the product. The more intricate the system, the less viable full automation becomes, and the more important it is to define what "assistance" means in practice.
Your membership also unlocks:
