Larian Studios has spent the last few weeks in hot water, following an interview with studio CEO Swen Vincke, in which he enthused about generative AI and its use in the development of upcoming, highly-anticipated fantasy game, Divinity. Now, in a public Reddit thread, Vincke has backtracked on his enthusiasm, detailing a new approach to generative AI use at the studio.
Per Vincke, there will not be any GenAI art in Divinity – and in fact, to preserve the integrity of the studio’s art team, the technology won’t be used at any stage in the concept art process.
Larian Studios CEO on GenAI approach: quick links
Larian Studios CEO claims no GenAI will be used in the Divinity concept art process
‘I know there’s been a lot of discussion about us using AI tools as part of concept art exploration,’ Vincke said. ‘We already said this doesn’t mean the actual concept art is generated by AI, but we understand it created confusion.’
‘So, to ensure there is no room for doubt, we’ve decided to refrain from using GenAI tools during concept art development. That way, there can be no discussion about the origin of the art.’
Larian Studios is seemingly looking to avoid any allegations of GenAI art use, and to avoid the possibility of actual GenAI art making its way into the final version of Divinity.
In recent months, we’ve seen games including The Alters and Anno 117 criticised for releasing with obvious GenAI assets included in the final product, with both instances causing major backlash from a passionate game-playing audience.
It’s worth noting Larian Studios developed its own passionate audience on the strength of the human-led narrative and characters developed for Baldur’s Gate 3, and that much of the backlash facing the studio comes from this same audience, who are loudly and justifiably against GenAI usage, for its erasure of human creativity.
But Larian Studios isn’t quite done with GenAI
Despite the loud outcry against Larian Studios’ admission of using GenAI to help develop Divinity, it does appear the studio isn’t quite discounting the technology yet. Further into Vincke’s Reddit thread, he described GenAI as being a potentially useful tool for speeding up other game development processes.
‘We continuously try to improve the speed with which we can try things out,’ Vincke said. ‘The more iterations we can do, the better in general, the gameplay is.’
‘We think GenAI can help with this and so we’re trying things out across departments. Our hope is that it can aid us to refine ideas faster, leading to a more focused development cycle, less waste, and ultimately, a higher-quality game.’
Notably, Vincke claimed the team would only generate ‘creative assets’ when they were 100% sure of the origins of the training data, and the consent given for that data to be used.
‘If we use a GenAI model to create in-game assets, then it’ll be trained on data we own,’ he said.
What’s notable about this statement is that proper training data relies on millions of data points to create coherent output. Training a model purely on a studio’s assets is unlikely to create the desired result, suggesting any GenAI output will need to rely on larger data sets, where consent is a more nebulous concept.
It means it’s unclear how exactly Larian Studios will effectively implement GenAI in its concept development processes, and whether it will actually serve development overall. For now, it appears the studio is still keen to test the waters, to see where GenAI could improve workflow and efficiency without treading on development disciplines.
