- cross-posted to:
- gaming@lemmy.ml
- cross-posted to:
- gaming@lemmy.ml
Excerpts:
The Verbal Verdict demo drops me into an interrogation room with basic facts about the case to my left, and on the other side of a glass window are three suspects I can call one at a time for questioning. There are no prompts or briefings—I just have to start asking questions, either by typing them or speaking them into a microphone
The responses are mostly natural, and at times add just a bit more information for me to follow up on.
Mostly. Sometimes, the AI goes entirely off the rails and starts typing gibberish
There are, of course, still many limitations to this implementation of an LLM in a game. Kristelijn said that they are using a pretty “censored” model, and also adding their own restrictions, to make sure the LLM doesn’t say anything harmful. It also makes what should be a very small game much larger (the demo is more than 7GB), because it runs the model locally on your machine. Kristelijn said that running the model locally helps Savanna Developments with privacy concerns. If the LLM runs locally it doesn’t have to see or handle what players are typing. And it also is better for game preservation because if the game doesn’t need to connect to an online server it can keep running even if Savanna Developments shuts down.
it’s pretty hard to “write” different voices for them. They all kind of speak similarly. One character in the full version of the game, for example, speaks in short sentences to convey a certain attitude, but that doesn’t come close to the characterization you’d see in a game like L.A. Noire, where character dialogue is meticulously written to convey personality.
I’ve played Ace Attorney and the writers put a lot of love and personality into the characters. I’d be sceptical if an AI could get close enough to any kind of writing style to “kill” writing in games like that.
Honestly getting fed up of AI doing a mediocre job of creating art and then people claiming it kills whole industries because it’s the “in” technology.
Yeah, the whole study of character is kind of outside the realm of an LLM. They can sort of reguritate, but they don’t think about humans like a humans do, the things we find funny or interesting about eachother, the small nuances that potential reveal a person’s history and internal experience.
An LLM, as sophisticated a piece of technology as it is, cannot really do that like a human writer can. Best it can do is sort of mimic in a limited capacity.
The amount of time to build something like this seems like it would offset the amount of effort it would take just to write good character dialogue. AI tools are basically word calculators, which means you have to provide data for the LLM, which means time to produce this data, time to build guardrails, etc. Even in this implementation, they say they had to build guardrails so that they don’t say anything “harmful.”
There are also a number of lawsuits going on that will set a precedent for how training data can be utilized in commercial products. While I expect them to take the side of large corporations with vast resources at the expense of ethics, there’s the possibility that they will do the right thing. This will affect how AI tools wil be used in such contexts.
I know the situation is different for everyone diagnosed with autism, but I like to compare AI writing as being something similar to someone with autism writing (as someone with autism). It can look kinda emotionless and robotic at times but other times it looks passable as something slightly less robotic.