Is there an irony in AI-powered game preservation?
I’ve been watching the recent explosion of game decompilation, recompilation and unofficial PC port projects with a lot of interest.
Games that were once effectively locked to ageing hardware are suddenly becoming playable in native form on modern PCs, often with higher resolutions, better framerates and all the other conveniences you would expect from a contemporary release. I’ve played the unofficial PC port of P.T. myself, and recently saw a video about an ICO PC port that sent me down this particular train of thought.
There is something fascinating about watching these games escape the hardware they were originally tied to.
ICO especially struck me because it is one of those games people routinely talk about as a piece of art rather than simply a piece of entertainment software. It has such a specific atmosphere and sense of authorship that preserving access to it feels culturally worthwhile in a way that goes beyond making an old game easier to play.
And now, through increasingly sophisticated technical work, that art can survive independently of the machine it was originally built for.
We are entering a period where old games are becoming more portable, more repairable and more accessible than they have ever been. AI-assisted tooling is accelerating efforts like decompilation, static recompilation, and reverse engineering, making it possible to preserve games in ways that would have sounded unrealistic not that long ago.
The strange part is that AI is kind of the hero and villain of this story.
On one side, it can help us preserve art.
On the other, it may fundamentally change the way we feel about the art being made.
AI makes it dramatically easier to recover old games, translate them, rebuild missing components, clean up assets, understand obscure codebases and make works that were once bound to ageing hardware accessible to almost anyone.
That feels like an obviously good use of the technology.
But the same underlying capabilities are also pushing towards a future where more and more of the content inside new games can be generated rather than deliberately authored.
Textures, dialogue, animation, concept art, voices, music, environments, even entire gameplay systems could increasingly be produced with some degree of automation.
I wonder what that does to the way we value games.
Part of what makes something like ICO feel meaningful is the sense that somebody wanted it to be exactly the way it is. Even the strange or imperfect parts feel intentional because they belong to a specific creative object made by specific people under specific constraints.
The limitation is part of the work.
If a game contains thousands of assets that exist because a model produced them cheaply and quickly, does that change our relationship with it? Even if the end result is technically impressive, do we instinctively value it less because we know less human attention went into making it?
There's always a danger in romanticising older games too much. Games have always been commercial products. Teams have always reused assets, followed trends, cut corners and worked within budgets. Not every texture in ICO was handcrafted because somebody believed passionately in the artistic significance of that particular wall.
But there was still scarcity.
Every asset cost time. Every environment needed people. Every line of dialogue had to be written, recorded and implemented. Creative decisions had friction attached to them.
AI potentially removes a lot of that friction.
That could be enormously liberating. We're already seeing small teams making things previously reserved for studios with hundreds of employees. Artists can iterate faster. Developers can spend less time on repetitive production work and more time on ideas.
But friction can also create meaning. If you can generate another hundred environments, another thousand NPC conversations or another ten hours of content almost instantly, what makes any of it important?
I also wonder whether the relationship between players and games has already changed before AI properly enters the picture.
A lot of modern games are not really presented as finished creative works in the traditional sense. They are platforms. Services. Seasons. Economies.
They have battle passes, rotating stores, engagement loops, daily challenges, roadmaps and content drops. They are designed not just to be experienced but to retain you.
For a younger player who has grown up with Fortnite, Roblox, Genshin Impact, Call of Duty seasons and constantly evolving online games, maybe the idea of a game as a fixed artistic object already feels old-fashioned.
Perhaps fewer people look at a game and think: this is a finished work made by a particular group of people at a particular moment in time.
Perhaps they now think: this is a place I spend time.
If a game is already understood as an endlessly changing service, does it really matter whether the next skin, quest, environment or voice line was entirely handcrafted?
Maybe all players will care about is whether it's entertaining, whether their friends are there, and whether the game keeps giving them something new.
And maybe that's okay? But there is still a slightly unsettling irony about using this technology to preserve old games because we recognise that they are culturally valuable, at the same time knowing that technology could make the games of the future feel more disposable.