Stellaris was launched in 2016., since then it has followed the usual Paradox post-launch cycle of regular tweaking with plenty of paid-for DLC expansions. is the latest The Machine Age and focuses on “artificial ascension” and the promise of eternal life among the stars, adding various new elements to the machine side of the game with a synth space queen who says she's going to take us all there. .
DLC's store page spells it all out then, at the very end, drops this disclaimer:
“We use creative AI technologies during the creation of some assets. This usually involves the concept of content and visual references. These elements represent a minor component of the overall development. is used to generate sounds for.”
Well, well, some Stellaris players say, do my eyes deceive me? Using AI creation tools in an expansion of the AI consciousness that governs the galaxy? In particular, the use of AI to generate sounds in the game raised some eyebrows.
It's a fair cop, says the game's director, before explaining how and why the team is using AI to help develop Machine Age. Stephen 'Aldrin' Morey was making piecemeal comments in Steam discussions before moving to the game's subreddit (Seen by RPS.) to offer more assurance to players.
“The AI voice generation tools we use at Stellaris ensure that the voice actors who sign up and create models receive royalties for every line we create,” says Morey. “Ethical use of AI technology is very important to us—we're pretty good at spotting dystopian sci-fi and don't want to go there ourselves.”
Morey promises to put together a more extensive post about how the devs use AI in a couple of weeks, but added that “we haven't used it for concept art in Machine Age. Because “there are some amazing concept artists on staff.” Among the other themed pieces are “visdev exploration/some AI generated pieces on the moodboard”, but these are being used as references rather than in any final art assets.
The director says he uses AI image generation tools to represent the basics of the ideas he's bouncing around with the game's designers, but as things progress “the artists then take our ideas and make the final May or may not use them to create assets. None of these design images go into play.” It's apparently the same with creative text, lots of “ideas” but “no conclusion or generated text goes into the game.”
Morey concludes by saying that Paradox has “strict guidelines” about using “AI tools legally and ethically” that “we abide by” by nature. It will certainly be interesting, when those giant diaries reveal their use, to learn more about them.
Overall, the approach the studio has taken seems to be very reasonable. Using AI to create image references in the conceptual stage is a step up from using Google Image Search, the more traditional way of putting together some kind of mood board, and until the final assets use AI imagery. is being done. The problem is hard to see.
The big problem for most seemed to be the two AI-generated voices, one of which is the player's advisor and the other is the DLC's synthetic Queen Satana. But the problem here is that the AI is creating those lines based on the work of human voice actors, who are then paid royalties based on their usage, and that the context is as appropriate as it can be. : An AI voice generator giving voice to a fictional AI. The character seems like a reasonable and really clever use of tech.
So I'd say Paradox comes out looking pretty decent. The times we live in mean that any game or expansion with this kind of AI disclaimer will be subject to a lot of scrutiny from players, who are heavily involved in the development process. remain skeptical about the unstoppable push of , but for most of the industry that ship has already sailed. Now the question is less about whether games made at a certain scale use AI, and more about how they use it, how any output is flagged, and whether that process I am replacing any of our meat bags or just part of a tool set.