The Divinity Developer Explains Its Implementation of Machine Learning for Next Divinity
The studio behind hit role-playing games like Baldur's Gate 3 and Divinity: Original Sin recently teased its new project, generating immense hype within the industry. However, follow-up statements from the studio's lead designer have brought a new dimension to the narrative, addressing the studio's approach toward machine learning.
AI as a Creative Assistant, Not a Substitute
In a new message, Swen Vincke outlined that the team is using generative AI for certain preliminary tasks. These encompass developing pitch decks, creating rough visual ideas, and drafting draft copy.
Importantly, Vincke stressed that the end content in the game will be authored solely by human writers. "We are creating everything in-house," he affirmed.
Our studio is continuously increasing our roster of concept artists and are actively putting together dedicated writer rooms.
As this area is being explicitly mentioned ā we currently have over twenty artistic staff and have job openings for more talent.
Everything we do is additive and focused on having people spend more time on the creative process.
Any ML tool applied correctly is supplementary to a developer's workflow, never a stand-in for their skill.
Responding to Feedback and Defining the Path
The revelation of AI usage initially sparked unease among a segment of the fanbase. In response, Vincke offered further detail on social media.
"At Larian, we employ these tools to explore references, similar to we use the internet and reference books," he stated. "In the conceptual ideation stages we use it as a basic framework for structure which we then substitute with original concept art."
He noted, "Larian brings on talent for their unique talent, not for their capacity to execute what a AI generates."
Key Areas of AI Integration
Vincke had earlier detailed the team's focused strategy to this technology, categorizing its use into key functions:
- Handling Monotonous Jobs: This encompasses refining animations, voice editing, and pipeline-specific tasks like adjusting assets for various species.
- Fast-Tracked Experimentation: Using technology to rapidly prototype rough mock-ups of gameplay ideas to validate concepts ahead of expensive production.
- Future Potential for Gameplay: Researching how machine learning could one day facilitate innovative reactivity, specifically in creating dynamic reactions in a vast role-playing world.
He specifically noted that core creative domains ā including writing ā are not areas where the team is cutting human input. Conversely, Larian is recruiting more in these very roles.
"We are neither shipping a game with machine-made assets, nor looking at trimming down staff to replace them with AI," Vincke concluded.
