'Nobody is Interested in Playing a Game That's Made by AI', Neil Newbon Says

Neil Newbon represented the Pixel Pack at the BAFTA Games Awards, offering up some critical home truths about AI in gaming.

[Christopher]

Gamers: "Companies are making crap games and putting no time into quality stories and gameplay."

Also Gamers (some of them the same people): "AI is the future, get with it."

[CrimsonWing69]

If it’s good, I would.

If a f*cking chimp made Baldur’s Gate 3 with the help of AI, why in the f*ck would I refuse to play it because of who or what made it!? I’ve played absolute dog sh*t made by human devs and pissed on all over by corporate greed, which in all honesty, could’ve probably benefited from AI.

Here’s something to contemplate: What program do devs use to place foliage in environments and what system does it use to generate the variations? Or what about Procedurally generated content? I’ll give you a hint, it’s AI.

[Christopher]

***If a f*cking chimp made Baldur’s Gate 3 with the help of AI, why in the f*ck would I refuse to play it because of who or what made it!?***

If thing that isn't happening and won't happen, then why not? What an argument.

***Or what about Procedurally generated content? I’ll give you a hint, it’s AI. ***

No, it's literally not AI. It's utilizing algorithms to generate foliage based on successful creations of foliage cooked into it. AI is a pattern learning method that can start with learning from algorithms but created pattern learned models for generation. It needs procedural content to learn patterns.If procedural is AI, then why is AI learning from procedural? It's not.

The fact you don't understand the difference tells me you don't understand AI or programming.

[CrimsonWing69]

*cough cough*

AI for level design can automatically place objects like trees and plants, making it way easier for developers to create complex and visually appealing environments and Gee willikers here’s some actual tools being used that I clearly don’t understand:

Promethean AI, which works with procedural pipelines.

PCGRL (Procedural Content Generation via Reinforcement Learning)
- Guess what this does, lol.

Houdini can be paired with AI scripts or Python plugins, letting developers bring machine learning into the mix to help guide and tweak procedural generation based on patterns the AI has learned.

There are AI tools like DreamzAR, iScape, and Fotor’s AI Landscape Generator that are already being used to design landscapes—complete with plants, trees, and layout suggestions. TREEEEEES with AI!!!

And my point with the chimp was, if a game is good, like on the level of Baldur’s Gate 3, and made by a chimp or with AI, why would I refuse to play it because of who or what made it? Can’t tell if you were picking up what I was laying down, there, buddy.

Your turn.

[Christopher]

SMH.

*** PCGRL (Procedural Content Generation via Reinforcement Learning) ***

WITH REINFORCEMENT LEARNING! It literally says it's procedural with AI on top of it. It is not just procedural generation.

Edit: Also, do you know how many things are calling themselves AI now that absolutely aren't AI? AI is the new marketing term for most people.

***And my point with the chimp was, if a game is good, like on the level of Baldur’s Gate 3***

Yeah, again, it's not happening. Stop bringing up science fiction movie ideas. This is reality.

[CrimsonWing69]

“***And my point with the chimp was, if a game is good, like on the level of Baldur’s Gate 3***

Yeah, again, it's not happening. Stop bringing up science fiction movie ideas. This is reality.“

I can’t tell if you’re trolling me with this one or you really don’t get what I’m saying…

And this one’s on me, while procedural generated content isn’t AI, the tools being implemented into modern game design to handle procedurally generated content once assets are made are AI solutions like Promethean AI.

So uh, yea. There ya go.

[Fishy Fingers]

I appreciate the social stance (jobs) as ive already seen it happening in my own professional field. But putting that aside, if you didnt know, or it was actually good, why would you care?

People will welcome what suits them and doesn't negatively affect them directly, whether thats AI cleaning my dishes typing up my meeting notes or making a game. Convenience, in the end trumps all.

[JEECE]

"People will welcome what suits them and doesn't negatively affect them directly"

This is it. As various events the last week or so have emphasized, the thing that really matters to people at the end of the day is the price they will pay for products and services as consumers. No one is going to go out of their way to play games made by/heavily reliant on AI if they are the same price as games that don't rely on AI and are worse quality. But if a company relying on AI makes a game people like, and especially if a company can make decent games and keep prices lower, gamers will buy their stuff.

Sure, there might be a small group of idealist gamers who refuse to play games where certain things are made by AI (I think voices will be a place certain people draw the line), just like some people will pay more for certain products because they are "fair trade" goods, but most won't care.

[Christopher]

***if you didnt know, or it was actually good, why would you care? ***

So far, the use of AI has been pretty blatantly obvious. From AI language tools that sound like really bad voice modulators to images in games or splash screens with typical AI generated issues. This argument of "was actually good" hasn't happened yet, especially not in video games. Generally, AI tools are for concept building in most game dev cycles, not entirely building games. The reason for that is because AI isn't as good as humans nor can it take direction.

[Fishy Fingers]

Agreed, with a massive caveat of 'yet'.

Hell, it probably does a great for analyzing web traffic, user activity and attention patterns but it probably can't do a great job of moderating a website, yet. But eventually, t'll be cheaper, constantly working, want less holiday, won't get sick... Will users care if its indistinguishable between Chris paying his mortgage or Chris the AI?

We should all get a trade :D

[SimpleSlave]

For now. Wait until some company comes up with the perfectly moronic AI made mechanic and makes bank with it, and then BOOM! Fuck the video game industry. Although, thanks to the nazi scum, we're already there, but still. Just like with DLC, incomplete games, season passes, patches, battle passes, loot boxes, horse armor, and digital only games, it's only a matter of time.

Good luck.

[Trilithon]

not sure about anybody else, but personally, im not supporting a game made by AI. i like my games made by humans and i am willing to pay them for it

[thorstein]

"AI" is not delivering what is being promised. It's a honeypot and people are going to lose billions (some already are).

The fake articles that proclaim its wonders are just advertising.

Will it continue to exist? Yes. But it's impact will be more akin to 3D printing (that didn't revolutionize manufacturing nor medicine).

[Christopher]

Also, everything has become AI even though it 100% isn't AI. Any data-centric automation system is suddenly AI and learning from things when it 100% is not doing that at all and it just relational database elements being referenced for actions in software.

[MrDead]

Have to agree, ai is costing far more than it makes, adoption has been slow and those that have adopted it are running into endless issues.

https://techhq.com/2025/03/...

"AI businesses run at a loss
The devastating side of these figures is that OpenAI, and we assume, its competition, is running at a tremendous financial loss. In 2024, the company accrued $4 billion in revenues, but cost $9 billion to run. That’s taking into consideration the 75% discount Microsoft gives its sister company on its AWS costs – hosting, compute, storage, load balancing, etc. If OpenAI were charged market rates for the substantial cloud services it uses, its costs would be $20 billion per year, making an effective loss of $16 billion a year.

According to The Information [subscription wall], training a model costs $3 billion alone, although Chinese DeepSeek claims it’s a whole lot cheaper. Even at DeepSeek’s cost levels, annual losses would still be immense. And so every evolutionary step of any significance comes with an eye-watering price tag.

At the heart of the AI bubble problem is that it costs big AI models more to come up with answers than people pay. The more users an AI platform has, the greater it costs the company running it, and the ratio of cost to revenue remains constant at any scale: every answer to a query is formulated on-the-fly.

On last year’s financial figures, if OpenAI were as popular as Gmail, its 10 billion monthly users would cost the company $56 billion a year. To cover that with subscriptions, at the same rate of conversion of free-to-paying as OpenAI advertises, each of the imaginary 96 million paying users would cough up $583 a year.

That would mean a near-tripling of the cost to use an AI model, assuming no one were to think the price too steep and opt out. The more users to decide against a tripled monthly charge, the higher the price tag to those that remain on the paying tiers.

And just to reiterate, those prices would allow OpenAI to break even – no profit for the company, and no return on the billions already invested by innumerable third-parties, including banks, venture capital firms, governments, and large corporations."

[Killer2020UK]

As a headline this is perfectly agreeable when it's implied the game would be wholly created by AI but the reality is not that cut and dry. Game development will almost certainly utilise greater amounts of AI to speed up development it's just that a human will almost certainly be providing the prompts and deciding what results are desirable or discarded.

I absolutely appreciate the ethical issues with using copywrited or existing content for datasets and then using those datasets to create models that effectively take the jobs of the people who created that original content. It just feels very much like a group with shovels vs one man in an excavator. Technology is here and things will inevitably move onwards.

[Knushwood Butt]

People will play whatever holds their interest, regardless of who or what made it.

People pay money for those jumping taco games.

When I take a glance at what my son is looking at on his phone, it's usually brain rotting stuff that runs for a few seconds and is then forgotten forever. You think he cares if AI is involved?

[Abear21]

I’m not. AAA gaming or small indie games, all should have a concept that evolves and improves over the course of the games development. Amazing art and different ideas impact the end product in ways the concept phase of game design cannot foresee. So could AI really develop a game that’s a complete AAA package? Even if it could, I’d rather give my hard earned dollars to actual developers who are creative for the love of the games and not outsourcing their creativity to a machine.

[bunt-custardly]

Neil Newbon - soon to be effectively replaced by AI which won't dish out its opinion at the BAFTA's when the game it developed wins the "best AI developed game" category. 2027.

[andyo13]

Ai is a tool. It can be used to speed up things like creating variations in trees and buildings. But shouldn’t be used to make a game completely from scratch. But 20%? 30%? What if it’s used well to give every pedestrian in gta a unique voice, personality, or even whole life with work and a home to go to?

[xenz]

I agree with him, as long as we humans make better games. If AI suddenly makes far better games, I see no reason for why I shouldn’t be interested in playing them.

[SimpleDad]

I would happily play AI created game that is good vs bugged mess with 2 hours of cast roll of outsourced companies.