The Creators of Baldur's Gate 3 Details Its Use of Generative AI for Next Project

The developer behind hit role-playing games like Baldur's Gate 3 and Divinity: Original Sin just teased its new project, sparking a wave of hype within the gaming community. However, subsequent statements from the company's co-founder have introduced clarity to the narrative, addressing the team's stance toward AI tools.

A Tool for Ideation, Not Replacement

In a new statement, Swen Vincke explained that the company is employing machine learning for certain ancillary functions. These include developing PowerPoint slides, generating rough visual ideas, and drafting temporary copy.

Crucially, Vincke stressed that the final material in the game will be created solely by human writers. "We are developing everything ourselves," he said.

We are actively increasing our roster of storytellers and are actively assembling writing teams.

As concept art is being explicitly referenced — we currently have twenty-three artistic staff and have job openings for additional artists.

All our efforts we do is supplementary and focused on letting our team spend greater focus on actual creation.

Every AI system used well is a boost to a developer's workflow, not a replacement for their skill.

Addressing Concerns and Clarifying the Vision

The admission of employing this technology originally sparked concern among some the community. In reply, Vincke provided more elaboration on social media.

"We use machine learning to research ideas, just like we use search engines and physical media," he wrote. "In the conceptual brainstorming phase we use it as a rough outline for composition which we then swap out with authentic concept art."

He noted, "Larian brings on talent for their inherent skill, not for their willingness to follow what a algorithm proposes."

Focused Uses for Machine Learning

Vincke had in the past detailed the team's focused method to machine learning, grouping its use into three main areas:

• Handling Monotonous Jobs: Areas like polishing mocap data, dialogue cleanup, and technical processes like adapting animations for different models.
• Accelerated Iteration: Using systems to quickly build rough models of mechanics to experiment with concepts ahead of expensive implementation.
• Long-Term Aspirations: Researching how machine learning could in the future create innovative gameplay, especially in simulating player-driven narratives in a detailed game universe.

He specifically noted that core creative domains — such as writing — are not areas where the team is cutting artistic talent. Conversely, Larian is recruiting more in these precise roles.

"We are neither releasing a game with machine-made assets, nor looking at reducing staff to swap them out with artificial intelligence," Vincke summarized.