Gamers Are Overwhelmingly Negative About Gen AI in Video Games | Hacker News Reader

Gamers Are Overwhelmingly Negative About Gen AI in Video Games

quanticfoundry.com

39 pointsjaredcwhite56 comments

Gamers are overwhelmingly negative of using Gen AI to replace what artists and designers would normally do. It feels artificial. What they could try doing instead is to use Gen AI for dynamic content generation during gameplay, like how Minecraft generates chunks. But more deliberate and intelligent rather than purely an algorithm. Gen AI is good for replacing not what humans would normally do, but what algorithms would normally handle but aren't great at.

I've always though the current AI technology we have would be perfect for generating new planets or fauna in a game like No Mans Sky. I feel most players wouldn't have a problem with it being used that way.

No Man's Sky already works without gen AI to do this incredibly well. I don't see what value gen AI adds when the current system allows for better tuning of parameters for generation and a gen AI model is more of a black box.

No man's sky generates nearly infinite variety but very little novelty.

What content generation would not fall under “what artists and designers normally do”?

When it's built into the game engine and is used as part of the game loop, responding to user input in some way, and evolving as the game unfolds. (As opposed to just using AI as a tool in the game creation process.)

Pretty sure whoever provides a fire and forget auto-UV unwrapper will have a statue erected in their image.

>Gen AI is good for replacing not what humans would normally do, but what algorithms would normally handle but aren't great at.

sadly, the 9000 layoffs in NA this year seems to show what they really want to do with AI.

On the art part at least, there's an analogy to Unity. No one liked low effort asset-flip games, and they gave Unity a bad name because they all showed the logo on startup due to using the cheapest license. But no one noticed or complained when good games were made with Unity, or even when they selectively used off the shelf assets. Similarly, we're going to go through a period where every terrible game is going to be full of assets with the stable-diffusion slop look, but no one will notice the AI-generated textures in good games that have actual standards for art.

On the other hand, I think I'd personally vote "somewhat negative" on quests and dialogue. Many games have too much pointless filler dialogue and unmemorable sidequests. Maybe if you don't care enough to actually write it, it shouldn't be there at all.

>no one will notice the AI-generated textures in good games that have actual standards for art.

I don't know. Clair Obscur is about as good as it gets and gamers notice. or at least, other devs who are also gamers notice it.

The community was always a much more "hyperactive" scene, and it doesn't seem like AI is an exception. Thing that may blow over in other industries will be eviscerated here, especially in a time where there's more cynicism than ever among the gaming community and the AAA industry.

True, I'm thinking more on the indie level. And "won't notice" is my hyperbole, of course sharp eyed people who use games and software are always noticing small technical choices even when you forgot them yourself. But it shouldn't dominate the impression of the product unless it's poorly done.

>But it shouldn't dominate the impression of the product

Well we're not exactly in normal times in the industry nor society at large. Deserved or not, the industry is in a hypervigilant state right now as a sense of preservation. We're even seeing serious unionization for once, that's how far it's been pushed back.

Stated preference vs. revealed preference: https://steamdb.info/app/1808500/charts/

Conflating actions and feelings is how businesses become hated. I'm not buying the expensive eggs because I believe they are a good price.

There’s a section “The Specific Use of Gen AI Matters” in the article. Arc Raiders uses GenAI to voice npc characters. I wish it didn’t, but this is not something that’d affect gameplay much.

AFAIK the enemies are also trained with machine learning.

That just shows that GenAI is associated with poor quality content at this moment.

So the true opposition is to poor quality content, not GenAI.

Unfortunately for artists, actors, etc. GenAI right now is the worst it's ever going to be, and it was much worse just a year ago.

Fortunately, GenAI will never be as good as human artists.

That may be true, but you can’t compare average GenAI with the best humans because there are many reasons the human output is low quality: budget, timelines, oversights, not having the best artists, etc. Very few games use the best human artists for everything.

Same with programming. The best humans write better code than Codex, but the awful government portals and enterprise apps you’re using today were also written by humans.

He didn't say "best humans". he said "humans".

And yes, I think with the current trajectory, we're not even hitting "crappy entrprise government portal" levels within the next few decades.

"Fortunately, machines will never be as good as human chess players."

Subjective vs objective. Also, analogies are almost always weak rhetorical distractions. The conversation just becomes about the differences between the two things. If you want to state an opinion about X, form the thought about X, rather than just pointing to Y and asserting they're the same.

> Subjective vs objective.

We've already reached the point where GAI art is extremely difficult to distinguish from human art and at a fraction of that cost.

I'd say that's pretty objective and it's hard to even leave room for subjective interpretation when it's so hard to tell them apart.

> Also, analogies are almost always weak rhetorical distractions.

I wasn't trying to start a discussion with them. To say GAI will never be better than humans at art when we already know what we know today isn't a good faith logical argument, it's a tautological appeal to emotion.

>where GAI art is extremely difficult to distinguish from human art and at a fraction of that cost.

And this logic is why people don't understand how to make good game art. generating an 2d animation or real time 3d model that properly deforms is multiple magnitudes different from fooling some tiktok users with a static image in isolation. even composing a still scene will quickly reveal the lmitations of generating art for your visual novel.

Wielding a camera doesn't make you a cinematographer that can sell a movie. Generating a few realistic-ish images does not make you an artist that can sell a game.

This holds true

https://journals.sagepub.com/doi/10.1177/09567976231207095

> AI Hyperrealism: Why AI Faces Are Perceived as More Real Than Human Ones (Nov 2023)

It's like the "can't stand CGI in movies" of decades past. Now there's CGI in almost every single movie. It's so good we can't notice most of it. The opposition was just to how low the quality ceiling was at the time, not really the CGI usage itself.

The thing is that only came about when companies realized that the thing they really wanted (to cut labor and costs of production) was counter to how you really make "good CGI". Bad CGI is still bad and good CGI makes some of the most expensive films to date.

Meanwhile, AI markets itself almost exclusively on being a time and money saver. And more efficient workers, but industry actively opposes that in a day and age where they prefer to commoditize labor instead of invest in specialists. If it doesn't actually do neither, then it won't really serve a niche compared to CGI.

I've been hearing "but it will get better" for a lot of things in the industry for decades now. It's a bit hollow to be future thinking when the present is collapsing around us for several reasons in and outside the games industry.

Doesn’t seem like they surveyed whether use of AI would actually affect purchasing or playing decisions. I think your implication is correct that it mostly won’t.

That’s a little harder to disentangle for a work of art. There is only one and if you want to experience it, you have to take the good with the bad. Plenty of people may hate a particular actor, but will still watch a film featuring them.

Owing to network effects, games can be extra sticky because you want to consume the same media as your friends.

I would prefer a mushy human was responsible for the art I consume, but I am probably not going to boycott a game because it uses AI for some assets.

Complete lie.

ARC Raiders uses AI for generated voice or TTS; it's hardly in the same field as generative AI which is mass scraped from the internet and churned out of a matrix math black box. They paid voice actors for the training and likeness, and the AI is used for future flexibility, and the VAs know this and its part of the contract they sign

They also use AI for the robotic kinematics which looks absolutely fantastic. When you blow up robots they zoom around and correct themselves exactly how you would imagine an evil robot to behave

This is a shining example of moving the goalposts.

How. When most people talk about distain for AI they are specifically talking about the use of generative AI and the fact that its widely used to astroturf creatives. Not all AI is bad, and not all AI is gen AI.

Replacing voice actors is literally using ai instead of creative people.

They did not replace voice actors. They hired voice actors to train the TTS they use. Very different to say ripping someone's voice off and not paying them royalties, residuals or any form of compensation whatsoever.

I'm somewhat of an AI-hater - maybe a bit more lukewarm about it than others - but I thought gaming (RPGs specifically) was a perfect use case for AI just because of dynamic dialogue.

I think AI has the potential to shrink dev teams back down to what they were the the 90s, which would be great for coherency and experimentation. Once you get a big enough org going there seems to be too much design by committee all over the place.

AAA is never going back to 10 people in a garage making Sim City. Nor funding a team that does this. The market, tech, and expectations of a game are too drastic for that to be the norm ever again.

As for the indie scene, AI isn't going to solve the problems on why they don't sell well. It's not assets nor tech that holds the scene back at the moment.

Well that scenario I have in mind is basically extreme AI geometry up scaling. You make a blocky draft and AI makes it pretty without touching the gameplay or level design. That seems feasible, no?

Just a very, very small, VERY vocal minority. 99,9999% of gamers (I was considering putting in a seven't 9, but decided to err on the safe side) just judge the end result and could not give a rat's ass about how it was created.

This is a survey of gamers that shows the overwhelming majority claim to hate it. Do you think they are only sampling that minority?

No. I'd say they are polling mostly completely uninformed people on the interwebs that haven't given this more than 3 seconds of thought before they clicked an option and have almost 0% expertise in actually detecting AI assisted work when it is done well, but since there is 0 cost proposed just prefered 'human' over 'machine'. Basically the equivalent of the beauty contestant's "peace on earth".

It doesn’t take more than a few seconds of thought for me to decide I don’t want the latest LLM slop in creative products. Inasmuch as video games are creative products.

Anecdote from me. I’m in a video games slack channel with ~350 of my coworkers who know well what ai looks like and like video games. Everyone hates it. I’d love a permanent steam selection to hide generative ai.

Mostly polling the uninformed would suggest they are actually polling the majority.

Do you need to be informed to know what kind of things in a game you personally like? This isn't some comprehensive survey to determine "what is AI?" The questions are literally framed as "what's your attiude". AKA "how do you feel?".

Given how little time a modern game has to advertise itself (without millions in an ad campaign), those gut feelings are key to landing a sale.

It's basicaly this, combined with the fact that most gamers think that AI means "slop".

But AI has completely legit non-slop use in games: texture gen, 3D model gen, Suno for audio, programming,...

Can you provide some examples of games where AI was used for results that are as good as human work?

AI hasn't been good enough for those tasks until six months ago. And it takes two years to make a game, so they aren't out yet.

Alright, I can't wait to hear how we're 6 months away from good AI generated art in another 2 years. Then another 2 years after that.

Dunno about games but there was a recent interesting pop video - Fatboy Slim & The Rolling Stones - Satisfaction Skank - https://youtu.be/_c_V3oPCe-s Not so much as good as human as able to do different things. I could see that kind of thing in games.

https://www.expedition33.com/

I agree but the issue is that C-suite executives think that AI can replace these things rather than enhance and you end up with lower quality trash. Look at sites like etsy etc., it's been overwhelmed with AI trash that is worth 0 value to anyone

If everyone on the planet is a gamer (8.26 billion people), your "99.9999%" figure would mean that only 8260 gamers--in the entire world--care.

I'm one of those gamers. I'm pretty sure we can find more than 8259 others.

Well Ackchyually award of the day

Reply on news.ycombinator.com