Sony are experimenting with AI characters who'll speak to the player using OpenAI and Chat-GPT, according to leaked footage of Aloy from Horizon Forbidden West.
It was kinda neat when they did it with Cortana. But then they removed all the Halo references and swapped the original voice actress with a generic voice synth, but kept the name for some reason. So what was once a charming little digital toy turned into a half-baked AI assistant that doesn’t do what you want half the time.
I imagine game studios are out of ideas so they are pitching, “AI NPC’s” to investors. This would also save money on voice acting and writing. If Rockstar’s GTA VI does bad then I’m counting that as the final nail in the AAA coffin.
Interesting point, although I don’t see how you’d manage to run modern TTS (the models can get very large, and that’s per voice; as an example Parler-TTS’s mini model is 800Mb, the HQ model is 2.3Gb - for one voice) + a LLM for content synthesis on any personal hardware, console or not. The storage requirements alone would make that grossly infeasible.
“…and a great deal of patience as you wait for each NPC to formulate their replies. In the meantime, they’ll just be standing there looking at you with glassy eyes, smiling.”
Does anybody actually want that, ever?
I could see AI interactions being interesting in some types of proceduraly generated games but I think the novelty would wear off pretty quick
It was kinda neat when they did it with Cortana. But then they removed all the Halo references and swapped the original voice actress with a generic voice synth, but kept the name for some reason. So what was once a charming little digital toy turned into a half-baked AI assistant that doesn’t do what you want half the time.
I imagine game studios are out of ideas so they are pitching, “AI NPC’s” to investors. This would also save money on voice acting and writing. If Rockstar’s GTA VI does bad then I’m counting that as the final nail in the AAA coffin.
Interesting point, although I don’t see how you’d manage to run modern TTS (the models can get very large, and that’s per voice; as an example Parler-TTS’s mini model is 800Mb, the HQ model is 2.3Gb - for one voice) + a LLM for content synthesis on any personal hardware, console or not. The storage requirements alone would make that grossly infeasible.
“This game requires a constant online connection”
“…and a great deal of patience as you wait for each NPC to formulate their replies. In the meantime, they’ll just be standing there looking at you with glassy eyes, smiling.”
crickets