007: First Light’s art director Rasmus Poulsen says developer IO Interactive hasn’t used any generative AI on the upcoming James Bond game, noting that, aptly, “the thematics of Bond are often: beware of utopia.”
It’s worth giving some context to Poulsen here, who was answering a question on the topic as part of a recent interview with Eurogamer, which you can read much more from in our big 007: First Light hands-on preview. As I put it to Poulson, it was interesting to note First Light’s story touches on the intersection of artificial intelligence and defence companies as a key plot point, and to be speaking to an art director in particular, amid the continuing discussion of gen-AI’s use in game development.
“I think it’s funny that you mentioned that,” he replied, “because, of course, the thematics of Bond are often: beware of utopia, I would say. And utopia comes in many shapes and forms. And in that sense, there’s certainly some thematics there about these things that we are faced with currently.”
In terms of using gen-AI then, he continued: “No we haven’t. We haven’t worked with AI on the project, generative AI.” Poulsen described the decision as a “combined discussion between core execs in the studio.” Personally, he added, “I think it’s a large discussion – I’d rather not dive into the details, because it’s complicated.”
Poulsen’s reticence is understandable, given the often fraught discussions taking place around AI, particularly generative AI, in video games today. Recently a similar sentiment was echoed by Tomasz Tinc, co-founder of The Blood of Dawnwalker developer Rebel Wolves, during a Q&A where Eurogamer’s Bertie Purchese was present. Fellow co-founder Konrad Tomsazkiewicz, meanwhile, attempted to straddle the line.
“I will just say one thing, because the subject of AI was raised, and whenever AI is raised, there’s controversy,” Tinc said. “I wanted to make one thing absolutely clear: nothing that’s in The Blood of Dawnwalker was created using generative AI, nothing. People with their blood and flesh made this game from the beginning till the end. I just wanted to make this 1000 percent clear.”
Tomsazkiewicz, for his part, added “I think that companies should use AI but in a way which helps people to work, not replace the people. For example, we have our own QA team and they sometimes have the task that they need to, for example, go through the terrain and check if there are any holes in the terrain. In the same time, they could play the quest and tell me if they like the characters or if the gameplay loop is fun enough, or if sometimes the combat can be better or whatever.
“My approach is I feel that we should use AI to help our people to work and take from them these tasks which are annoying and frustrating and allow them to do this more fun work, which is needed, actually. And this is what I think about it. It should make our life easier but it shouldn’t replace people at all. It’s like you can take a stone to put the nail into the board or you can use the hammer – it should work like this.”
Meanwhile, Google Cloud’s global director for games Jack Buser made headlines in our sister site GamesIndustry.biz recently, with claims that AI could “right size” the “not sustainable business model” of game development, though notably Buser didn’t give specific examples of how that’s currently being done in practice. (Google is currently one of the major stakeholders in the success of gen-AI, via its Gemini suite of tools.)
To come back to Poulsen, all that makes for understandable reluctance to get too into the weeds on personal opinions about the topic. Nevertheless, this makes IO Interactive the latest developer, alongside the likes of Phantom Blade Zero’s S-Game, Dispatch’s AdHoc, Suda 51 of Grasshopper Manufacture and more, to make it clear it’s not using generative AI at all in development.
