Neowiz’s AI-focused job listing for the Lies of P sequel has sparked backlash. We break down what the role actually says, why it has fans on edge, and how generative AI could reshape art pipelines, trust, and the future of this rising Soulslike franchise.
A Sequel Enters Full Production, and an AI Job Ad Steals the Spotlight
Neowiz and Round8 have quietly moved the Lies of P sequel into full production. Inven Global reports that the studio is staffing up across disciplines while doubling down on narrative, with new hires aimed at strengthening scenario writing and direction for the next chapter in Krat.
That all should have been good news. Instead, the big talking point around the sequel this week is a very different kind of hire: an “AI creator” role built explicitly around generative AI tools.
A job listing surfaced on Neowiz’s careers page describing a position that would use systems like Stable Diffusion and Midjourney to generate character and background concept drafts, create textures and other in-game assets, and even train bespoke AI models that help define a project’s visual identity. Coverage from Push Square and PCGamesN highlighted the listing, and the backlash was immediate.
For a game that built its name on handcrafted, ornate horror, this has fans asking whether the sequel’s art will still be made by people at all.
What the AI Role Actually Describes
The wording of the ad is important, because it goes beyond vague “R&D” language.
According to the summaries of the listing, the AI creator would be responsible for:
Creating early visual drafts of characters and environments using image-generating models, then iterating on those prompts to explore directions quickly.
Turning AI output into usable game assets, including textures and possibly base meshes for 3D models.
Training internal, “unique” AI models tailored to the studio’s art style, suggesting Neowiz wants systems that can repeatedly generate on-brand imagery.
Collaborating with art teams to integrate these tools directly into the pipeline for future projects.
This is not framed as a pure tools engineer or research role. The posting calls the position an “AI creator” or “AI artist,” which is exactly the phrase many artists and players have pushed back against over the past two years.
Why Fans Are Pushing Back So Hard
Lies of P earned praise partly because it looked so richly deliberate. Krat’s Belle Époque streets, its twisted puppets and stained-glass blues, felt like they came from people obsessing over every flourish.
Putting that same series on a path where AI is explicitly tasked with “creating concept drafts” and “generating textures” hits several nerves at once.
First, there is the credit and consent issue. Most popular image models are trained on enormous datasets scraped from the web, including work from concept artists and illustrators who never agreed to have their styles ingested. Players who care about the craft of game art see an “AI artist” role as inherently built on uncredited labor.
Second, there is fear of displacement. Even if studios insist that AI is only for rough drafts or background assets, a dedicated role that makes AI imagery production its core responsibility signals a structural shift. If AI can deliver acceptable base concepts in minutes, management may eventually question why so many human concept artists are needed.
Third, Lies of P carries an unintentional irony. As some critics have pointed out, the original game can easily be read as an allegory about automation, control, and what is lost when machines override humanity. To see its publisher lean into generative AI so openly feels to some like missing the point of their own story.
Finally, there is the timing. Neowiz has been talking up the sequel’s expanded narrative ambitions and investment in storytelling, while the AI job ad makes headlines for automating parts of the visual identity. That contrast sends a mixed signal about where the studio thinks the “soul” of the series really lies.
How Generative AI Could Reshape the Art Pipeline
Stripping away the emotion for a moment, the role Neowiz describes fits neatly into how many studios are quietly experimenting with AI behind the scenes.
In pre-production, AI could be used to blast through dozens of visual explorations for a single boss or district in Krat. A human art director might prompt a model with references, pick a handful of interesting thumbnails, and then have concept artists paint over or refine those results, instead of crafting every initial sketch manually.
For environments, AI texture generators can output convincing stone, wood, cloth, or ornate metalwork patterns in bulk, which artists then adjust, tile, and combine in tools like Substance. The AI is effectively a noisy material library that replaces some of the grunt work of building base maps.
Studio-specific models take this further. If Round8 trains an internal system on approved Lies of P art, the model could become a specialized engine for “Lies of P style” imagery: puppet joints, twisted marionette silhouettes, decadent theaters, and gaslit alleys. That could speed up asset ideation dramatically, especially on a sequel that needs to feel familiar but fresh.
This is the optimistic read from a production standpoint. Used carefully, AI can reduce the time spent on low-level iteration and let senior artists focus on composition, storytelling details, and polish. Games do not get cheaper or faster to make, but some of the most repetitive tasks could shift from brushwork to curation and paint-over.
The risk is that economic pressure pushes that balance in the opposite direction. Once you have a pipeline where AI can output passable concept boards, it becomes easier for management to minimize expensive, time-consuming human passes. The more of Krat’s visual DNA that is learned and replicated by a model, the more tempting it is to let that model generate not just reference, but final background assets, props, or even marketing art.
The Trust Gap Between Fans and Pipelines
The backlash to Neowiz’s listing is not happening in a vacuum. Across the industry, studios have quietly been using generative tools for everything from mood boards to placeholder art, but players generally do not see that process. What sparks outrage is when a company centers AI in its public messaging or hiring.
In Lies of P’s case, the role is explicitly presented as a creative driver. For fans, that sets off alarm bells about the sequel’s authenticity. It is not only a question of whether AI will hurt jobs, but whether the game will feel like something made because someone cared, or something manufactured from a prompt archive.
Once that doubt takes root, it colors everything else. Trailer shots get scrutinized for telltale AI artifacts. Concept art is questioned for weird fingers or asymmetrical filigree. Fans start to wonder whether the studio is using the money they spent on the first game to replace the very artists whose work attracted them.
This erodes the fragile trust that underpins long-term support. Lies of P’s sequel will likely ask players to invest again, whether through a full-price purchase or early preorders. If a chunk of the audience is already promising to sit this one out because of AI, that is damage done before a single frame of gameplay has been shown.
What This Could Mean for the Sequel’s Look and Feel
The core worry is that the sequel could lose the sense of bespoke craftsmanship that defined Lies of P. Art direction in the first game was tightly controlled: there is a consistent, painterly haze to Krat’s lighting, and each weapon, coat, and mask has a lived-in texture that feels touched by human hands.
An AI-assisted pipeline risks diluting that cohesion if not managed carefully. Different prompts or model versions can produce slightly divergent styles, and if those are not rigorously harmonized by lead artists, the result can be a world that feels subtly disjointed. In a Soulslike, where atmosphere is as important as hitboxes, that can be fatal.
On the other hand, if Round8 is serious about training internal models on strictly curated, studio-created data, the sequel could actually lean deeper into its own aesthetic. Imagine a tool that instantly generates multiple variations of Lies of P’s puppet aristocrats or grotesque carnival contraptions, all filtered through the art bible the team has already established. With strong art direction, AI becomes a fast sketch assistant, not an invisible co-author.
The problem is that players have no direct way to verify which of these futures Neowiz is aiming for. Without clear messaging, fear defaults to the worst-case scenario: a sequel whose gothic opulence has been partially outsourced to a black box trained on other artists’ uncredited work.
Can Neowiz Rebuild Goodwill Before Release?
If Neowiz wants the conversation around the sequel to shift back to its ambitions rather than its tools, transparency will be key.
First, the studio could draw a firm line between AI and final assets. Outlining that AI is used only in early ideation, that every visible asset still passes through human hands, and that the core concept and character art teams are growing rather than shrinking would go a long way for skeptical fans.
Second, Neowiz could address the training data question directly. If internal models are built only on in-house work or licensed datasets, saying so upfront would distinguish the studio from the many companies quietly relying on broad, scraped corpora. That sort of commitment would be unusual, but Lies of P’s audience is exactly the group paying attention to that level of detail.
Third, the studio’s other messaging about the sequel needs to highlight its human creatives. Inven Global’s reporting on the focus on narrative hiring is a good start. Showcasing key artists and writers in dev diaries, concept breakdowns, and GDC talks would reinforce that people are still at the heart of Krat’s return.
None of this will please everyone. Some players will avoid any project that touches generative AI at all. But clear boundaries and visible respect for the craft can keep the conversation from collapsing into pure cynicism.
The Bigger Picture: Where This Leaves the Lies of P Franchise
Lies of P sits at a crossroads. The first game carved out a spot among modern Soulslikes by pairing firm combat design with a distinct visual identity and a surprisingly sharp narrative hook. The sequel is being built in a landscape where tools and expectations are shifting fast.
Strategically, Neowiz’s AI push suggests the publisher wants to scale up. The company is already involved in multiple PC and console projects, and internal AI models that can generate consistent art across games are attractive if the goal is to ship more titles without ballooning headcount.
For the franchise, that creates a tension between becoming a reliable, expanding property and preserving the handcrafted feel that made it special. If the sequel lands as a critical and commercial hit while using AI in its pipeline, it will be held up across the industry as proof that players will accept generative workflows, at least when the final product is strong.
If it stumbles, though, every awkward animation, stiff-faced NPC, or oddly patterned backdrop will be blamed on AI, whether that is fair or not. Lies of P could become the case study people cite when arguing that you cannot prompt-engineer your way to soul.
In that sense, the sequel is not just another Soulslike follow-up. It is one of the first high-profile tests of how a mid-sized, artist-forward studio navigates generative AI in public. Neowiz’s hiring push has already shown that players are watching closely. How they respond from here may shape not just Krat’s future, but how comfortable other studios feel about putting AI at the center of their own next big thing.