Industry Insider Says AI Cuts Game Art Costs by 50% — A 10-Person Team Went Full AI and Raked in Hundreds of Millions, and Artists Are Losing It

A 10-person dev team went full AI on art, raked in hundreds of millions in its first month, and went toe-to-toe with a massively budgeted 3D title — and when that story collided with industry reports claiming AI can cut art production costs by 50%, the entire game art community collectively lost its mind.

According to industry reports making the rounds on NGA (China's largest gaming forum), insiders say game development traditionally demands designers with 3-5 years of experience for art roles. But with AI tools, even junior designers can now produce solid results, driving down HR costs significantly. The headline figure? Art departments could see up to a 50% cost reduction. To put that in perspective: a 20-person art team could potentially be replaced by 10.

The top reply cut straight to the chase: "Does this mean gacha games that survive purely on character art can become even cheaper to make? Companies like [a certain publisher] that basically live off 2D illustrations could see even fatter profit margins." Pure waifu collector games are the biggest winners of AI cost-cutting — their already-thin art budgets could get slashed in half.

Reply #2 dropped names directly: "Some games are already using AI for character art — like Rayark, that AI-happy studio." And Reply #3 delivered the brutal reality check everyone was thinking: "Costs went down, prices stayed the same, they're just pocketing the difference." Players haven't seen a single discount despite all the AI efficiency gains.

But Reply #4, from someone claiming to be in the trenches, offered a reality check on the hype: "The future is promising, but current capabilities aren't that powerful yet. Moron managers keep screaming 'use AI,' but in practice, doing it manually is faster for most tasks. AI is decent for quickly cranking out reference images though." In real production environments, AI is more of a reference tool than an actual replacement — for now.

The most insightful comment came from Reply #9, a lengthy essay that became this thread's most celebrated take. They compared AI-generated waifus to pre-made microwave meals: "Anyone can buy a frozen meal from the supermarket — when you go to a restaurant, it's because you want a real chef cooking your food fresh." If AI waifus were so beloved by players, why does every studio accused of using AI immediately panic-post stacks of rough sketches to prove human involvement, instead of proudly owning it? They argued that AI batch generation is something literally anyone can do — so why would players pay for characters with "low human content"? The comparison between creative assistive tools and AI cost-cutting was razor-sharp: "The former is bringing a calculator to a math exam. The latter is copying answers from a completed test paper."

Reply #15 pushed back with a more nuanced take on how AI is actually being used in practice — not the "generate and polish" approach that's widely mocked, but as a front-end design accelerator. "Humans can only produce 2-3 design iterations with huge time investment. AI can crank out dozens or even hundreds in the same time. The remaining work is human curation and manual stitching. Once the design is locked in, it's back to skilled artists grinding out the final product." In other words, AI art IS human art — sort of — with unlimited rough drafts available, but quality depends entirely on the person picking and refining.

Reply #16 confirmed the pain from a freelance artist's perspective: "My friend took an illustration commission where the client had already run their sketch through AI to generate a character art piece and said 'just fix this up.' My friend spent hours cleaning up the AI output and swore never to take that kind of job again. It's absolute torture." Touching up AI-generated art was described as an excruciating workflow experience.

The real bombshell in this thread was the case of 'World Beyond' (世界之外), NetEase's otome game mentioned in Replies #10 and #18. Reportedly running on deep AI involvement with just a 10-person team, the game leveraged marketing buzz from riding the coattail of Papergames' (叠纸) new release to go straight into a public launch. Despite minimal development costs, it reportedly hit 5th place on the charts with first-month revenue in the hundreds of millions — going blow-for-blow with Papergames' massively expensive 3D otome rival 'Love and Deepspace' (恋与深空). Reply #18 called this the "AI dividend period" — an absurdly high ROI.

Palworld was also dragged into the discussion as further proof of AI efficiency gains — with its studio head reportedly claiming their art production was 5x faster than normal. "High EQ: incredible efficiency. Low EQ: AI went brrrrrr with the copy-paste."

Reply #17 from someone with insider connections painted the grim human cost: "People around me have been laid off because of this — just a speck of dust from the times falling on individuals. Every major studio is using AI to varying degrees, from simple reference images to running roughs through AI for workhorses to polish, all the way to straight-up generating from other people's art. A senior art director's assessment: it's reached the level of a 10+ year veteran illustrator." The silver lining, they noted, is that top-tier artists are still largely unaffected — past a certain skill threshold, drawing by hand is actually faster. But "the future is hard to say."

Reply #12 raised an even more unsettling question: if World Beyond went full AI on art, does that mean the writing was AI-assisted too? If art can be automated, why not scripts?

As for why players despise AI waifus so passionately, Reply #9 already nailed it: the undifferentiated human labor embedded in each image has decreased. When players spend real money and time for their beloved characters, they naturally want those characters to have been crafted stroke by stroke by a human artist — not spit out by a batch generator.

The bottom line: AI cost-cutting in gaming isn't a future prediction — it's already happening. The only question is how deep it goes. When a 10-person team can generate hundreds of millions in revenue, every studio still running traditional workflows gets pressured to follow suit. And players have made their stance crystal clear: cut costs all you want, but don't let those savings come at the expense of their gaming experience. As for studios trying to use AI on the sly while dodging player scrutiny — they're destined to keep getting caught when time pressure forces sloppy AI touch-ups that millions of eagle-eyed players will spot instantly.