CD Projekt RED is testing AI tools to build livelier towns and streets. AI NPC generation for Witcher 4 aims to automate prototyping, speed up QA, and create dense, believable crowds with varied looks and routines. Similar research supports Cyberpunk 2, promising richer worlds without replacing writers or actors. CD Projekt RED’s latest report confirms an internal AI team working on tools that help build and test game systems faster. The group has explored automation for prototypes, quality checks, and large-scale NPC crowds across The Witcher 4, Cyberpunk 2, and the online-focused Project Sirius. AI NPC generation for Witcher 4 is part of this push, but it supports developers rather than replaces them.

How AI NPC generation for Witcher 4 could work

Procedural variety that avoids copy-paste crowds

Crowds often break immersion when they reuse the same faces, clothes, and movements. AI-assisted pipelines can mix assets and animations at scale so each passerby feels distinct.

Blend clothing, body types, and accessories to reduce clones

Swap micro-animations like fidgeting or idle glances

Attach light backstories for believable barks and actions

This approach to AI NPC generation for Witcher 4 is about scale and speed. Artists still set the style and rules, while tools fill in safe variations to populate markets, inns, and city gates.

Schedules, reactions, and crowd flow

Smart NPCs follow simple daily plans and react to the world. AI tools can help generate and test those routines faster.

Daily loops: work, shop, rest, travel between districts

Local rules: guards check trouble spots; vendors respond to time and weather

Event triggers: crowds gather for fights, flee from danger, or cheer bards

By simulating hundreds of agents, designers can spot pathing jams and tune crowd density. This supports smoother movement, fewer pileups, and more natural reactions.

Faster prototyping and smarter testing

The report highlights automation for prototyping and QA. In practice, that can mean bots that stress-test maps, measure frame times, or try weird player actions.

Automated path checks to find stuck spots before playtests

Performance sweeps to see how many NPCs a scene can hold

Behavior audits that flag broken schedules or looping states

The result is more time for human designers to craft quests, write characters, and polish set pieces.

Shared wins for Cyberpunk 2 and Project Sirius

CD Projekt RED says its AI research spans multiple games. Lessons from one setting can benefit another.

Cyberpunk 2 can use improved traffic and pedestrian flow for denser streets

Project Sirius, with online elements, can scale crowds and events more reliably

Tools built once can reduce rework across future updates

The studio notes it has about a dozen AI research tracks running, with some already in testing. That suggests steady, practical adoption rather than a single big leap.

Tech fit: DLSS 4/5 and RTX Mega Geometry

The Witcher 4 is set to support Nvidia DLSS 4, with possible DLSS 5 later, plus RTX Mega Geometry. These features can free up performance that crowds usually eat.

DLSS can upscale frames to keep high detail in busy scenes

RTX Mega Geometry can handle rich meshes with lower VRAM cost

Better budgets allow more NPCs on screen without harsh frame drops

When paired with smarter crowd logic, these gains can turn “packed but choppy” plazas into smooth, atmospheric hubs.

Creative lines developers say they will not cross

Many studios, like Capcom, say they avoid using generative AI for final game content but will use it for development. CD Projekt RED’s report echoes this idea: research targets workflows such as prototyping, testing, and crowd creation. That does not mean replacing writers, actors, or quest designers. Clear boundaries help protect story quality, voice talent, and player trust. Ethical use also means reviewing data sources, reducing bias in crowd models, and keeping human sign-off on anything that reaches players.

Release timing and expectations

The Witcher 4 is not planned for 2026. Current estimates point to 2027 or later. Until then, expect slow, visible progress: better tools, cleaner builds, and occasional tech details. If the studio hits its goals, players should see:

Busier markets and roads that feel alive

Fewer look-alike NPCs and stiffer loops

More emergent moments that fit the setting

Smoother performance in crowded spaces

Strong tools rarely grab headlines, but they set the floor for the kind of worlds CD Projekt RED wants to build next. In short, AI NPC generation for Witcher 4 looks like a behind-the-scenes upgrade that helps teams ship richer crowds, smarter routines, and steadier frames, while keeping human creators in charge of the heart of the game. (p (Source: https://www.pcguide.com/news/cd-projekt-red-has-been-testing-out-ai-tools-to-create-realistic-npcs-in-the-witcher-4-and-cyberpunk-2/)

For more news: Click Here

FAQ