What AI Is Really Doing to Games and the Game Industry
Written by Michail Katkoff, who uses AI daily to boost productivity, spark creativity, and occasionally generate ethically questionable spreadsheets.
The future of game development won’t be about speed. It’ll be about ethics, intent, and ownership.
In a recent conversation with George Ng, CTO of GGWP, I got to explore the real impact of AI. The conversation was less about what it can do and more about how we’ll live with it. And I think that’s a bit more interesting angle to discuss.
So much of the AI talk is sales pitches by technology overlords and super-funded AI startups.
Meta, Google, Microsoft, OpenAI… All take turns scaring CEOs around the world that if they don’t adopt AI, they will find themselves “blockbustered”. Living a “Kodak Moment” as their peers pass them by.
At the same time, where are those AI native companies that are passing the incumbents? The biggest success stories I’ve seen are kids' vibe-coding apps that make a few million. Good on them! But vibe-coding the next app for identifying your napping window or an app that measures social battery is not exactly revolutionary.
I wrote down my reflections based on the conversation with George.
1. AI Isn’t Creating New Problems. It’s Supercharging Old Ones
George was clear: AI doesn’t invent ethical issues, it scales them. Personalized pricing? Manipulative matchmaking? AI just makes these systems more efficient, and therefore more dangerous.
The tools now exist to run free-to-play’s dark patterns with surgical precision. Not just variable drop rates, but dynamically optimized offers based on churn prediction, social graph, and psychological profile. With AI, player manipulation becomes not just scalable, but elegant.
We started with A/B testing, expanded into segmentation, and the next level is personalization. The right difficulty and the right offers for the right player at the right time. Guaranteed lift in the players’ lifetime value.
The regulators are actively discussing the harms of algorithms in social media. Are the AI driven personalization algorithms the next target? Could be. Depends on the impact on the players’ credit card bills.
2. Ethical AI Will Be the Differentiator
In our conversation, George flagged the growing performance gap in moderation tools trained predominantly on Western datasets. Thai speakers? Women? Kids in mature-rated games? The models simply don’t get them. This isn’t evil. But it is structural bias, and it will quietly shape who gets muted, who gets banned, and who gets heard.
The unspoken reality: no one is really accountable, yet. Studios deploy AI with a shrug, “we’re just using off-the-shelf tools.” But as these systems make more decisions, from moderation to monetization, the accountability vacuum becomes a commercial and legal risk.
The next gold standard? Not just faster AI. But auditable, explainable, and equitable AI. In the context of gaming, this means building AI systems that treat all players fairly regardless of language, location, age, gender, or socio-economic background whether the tools are for moderation, personalization, matchmaking, or content generation.
3. The Entry-Level Pathway Is Breaking. And Nobody Has a Fix
I wrote back in September: “AI replaces process, not people. But what happens when people are in the process?” George took that further, pointing to the eroding middle tier of studio hierarchies. Junior devs are now AI-assisted. Seniors are reviewers and owners. But the mid-level layer, the human throughput of the studio machine, is thinning.
There’s a correlation between AI adoption and graduate unemployment rates. Recent US graduates have seen a much sharper spike in joblessness, roughly doubling their unemployment rate to 5.8%. Not to mention that a significant share of new grads remain stuck in under-skilled roles, with underemployment reaching over 4%.
This creates a weird paradox: more tools, fewer apprentices. We’ll see smaller teams ship bigger products faster, yes. But we may also see fewer new seniors emerge. Where exactly do they get trained if the entry and mid-level positions get cut? And what does to future look like when the funnel from an apprentice to a master is non-existent? Or is this an obsolete way of thinking altogether? Because in the AI era, building have become democratized as seemingly anyone can code their idea into software.
4. AI Companions Are Already Real Enough to Hurt You
Replica users are experiencing heartbreak. NPCs with memory. Dating sims are becoming therapists. This isn’t speculative fiction. It’s user behavior. George recalled cases of players feeling abandoned or betrayed when AI companions changed tone after model updates or content policy adjustments.
These systems already trigger real emotions. Whether they deserve those emotions is irrelevant. What matters is that players feel something, and that games are now explicitly optimizing for that kind of attachment.
What’s the monetization model for synthetic intimacy? What happens when game developers begin exploiting the loneliness epidemic with AI NPCs? If you thought creating a feeling of loss aversion and monetizing on it was bad, wait till some games adopt straight-up webcam monetization strategies.
5. We’re Going to Drown in Good-Enough
AI will supercharge reskins, clones, and vibe-coded shovelware. It already does. The barrier to launch is now measured in hours. Puzzle levels? Slot reels? Narrative branches? AI can output thousands.
But here’s the catch: most players already have their “forever game.” And most games already find themselves dead on arrival. AI won’t fix that. If anything, AI will amplify the power of the incumbents. The best IPs, with the biggest networks, the deepest UGC hooks, and the most trust, will dominate harder.
The same tools that let you build a game in a weekend also let you create something forgettable at industrial scale. In other words, when building becomes easy, distribution becomes the kingmaker.
Philosophical Musing
At the moment, I believe the promises of AI are exaggerated. It feels like the dot-com bubble all over again. In the late 1990s, the Internet was going to change everything. Pets.com, Webvan, eToys, and hundreds of other startups were absolutely right about the future. Just not about the timing.
Aream & Co’s recent CEO study revealed that 80% of the companies have limited implementation of AI. This is happening despite the tech companies scaring these executives daily to do something, anything, with the AI so that they are not “blockbustered”.
We’re watching the AI sector go through a familiar arc: massive capital inflows, frothy valuations, and a gold rush of companies promising that LLMs, generative media, and automation will rewrite the rules of every industry, especially gaming. And like the dot-com boom, it’s built on truths that are arriving ahead of their time.
That said, I don’t doubt that AI will bring massive change, just like the internet eventually did.
When it comes to gaming, I feel like AI’s grip on gaming is tightening. It’s doing so not with a bang, but with a quiet series of check-ins, patch updates, and dev tool plugins.
AI won’t remake gaming overnight. But it will change who gets to make games, how they’re made, and who they’re made for. The tools are neutral. The incentives aren’t. And in an industry this unregulated, the most important decisions will be made not in policy meetings, but in build reviews, GDDs, and Jira tickets.
The real question isn’t: “What can AI do?” It’s: “What are we going to let it do?”. Right now, we’re forcing AI into development without even pausing to ask what the implications might be.
