The Divinity Developer Details Its Implementation of Generative AI for Next Divinity Game

The studio behind acclaimed titles like Baldur's Gate 3 and Divinity: Original Sin recently shown its upcoming project, sparking a wave of anticipation within the gaming community. However, subsequent remarks from the company's lead designer have added a new dimension to the discussion, focusing on the team's philosophy toward machine learning.

A Tool for Ideation, Not Replacement

In a new clarification, Larian's director outlined that the team is utilizing generative AI for particular ancillary purposes. These include fleshing out presentation materials, generating rough concept art, and writing placeholder copy.

Crucially, Vincke emphasized that the end content in the game will be crafted entirely by real creatives. "Our team is developing every line in-house," he affirmed.

Our studio is continuously growing our team of writers and are busily forming narrative groups.

Since this area is being specifically called out — we presently have over twenty artistic staff and have positions available for more creatives.

Each initiative we do is supplementary and focused on enabling creatives to spend additional energy on the creative process.

Every machine learning application implemented properly is additive to a creative team process, never a stand-in for their talent.

Addressing Concerns and Clarifying the Vision

The news of AI usage initially sparked unease among portions of the player base. In reaction, Vincke issued additional elaboration on online platforms.

"At Larian, we employ machine learning to gather inspiration, similar to we use Google and art books," he wrote. "During the very early brainstorming phase we use it as a basic framework for layout which we then replace with hand-crafted concept art."

He noted, "We've hired talent for their unique talent, not for their capacity to replicate what a AI generates."

Key Areas of AI Integration

Vincke had previously outlined the studio's practical approach to machine learning, categorizing its use into three main pillars:

• Handling Monotonous Jobs: Areas like motion capture cleaning, voice editing, and technical processes like retargeting animations.
• Fast-Tracked Experimentation: Using systems to speedily create basic models of gameplay ideas to validate concepts ahead of expensive production.
• Future Potential for Gameplay: Investigating how machine learning could one day enhance innovative gameplay, particularly in simulating unforeseen permutations in a vast role-playing world.

He specifically affirmed that core creative areas — including writing — are not areas where the company is reducing artistic involvement. On the contrary, Larian is recruiting more in these precise roles.

"Larian is not shipping a game with AI-generated content, nor looking at cutting staff to swap them out with AI," Vincke concluded.