Everyone knows the moment when you get stuck in a game and somehow can’t make progress anymore. The question is whether the solution should be for the game to take over instead.
An unbeatable boss. A level that keeps ending in failure. In moments like these, it may not be the player who improves in the future, but the game itself that takes over.
The Ghost Within The Game
According to a report by VGC, Sony has filed a patent for so-called AI ghosts that describe new forms of in-game assistance. The patent was originally submitted in September 2024 but only drew wider attention this week after the World Intellectual Property Organization published an international status report on it. The AI ghosts outlined in the patent are intended to give players hints when they get stuck or, in some cases, allow an artificial intelligence to complete a challenge entirely on their behalf.
The AI ghosts would not simply be programmed to repeat specific button inputs, but would instead be trained on real gameplay. In doing so, they are meant to actually learn how a game works in order to provide players with more targeted assistance.
When Help Becomes A Shortcut
Regardless of that, the patent raises a fundamental question: what role should artificial intelligence play in game design going forward? It would continue a trend that has been running through gaming for years, making games increasingly easier and more accessible.
In principle, there is nothing wrong with making games appealing to a broader audience. Extensive quality-of-life features or clear tutorials are a benefit to almost any game. It becomes problematic, however, when a game’s central challenge is softened to the point where it can be overcome regardless of effort or willingness to learn. That is a design decision I can no longer support.
This is precisely where the concept of an AI assistant in video games becomes problematic. If an AI does not explain systems or help build skills, but instead simply solves tasks, the focus shifts. Gaming then no longer means overcoming a challenge, but merely moving forward.
To me, this stands in direct contradiction to what gaming is fundamentally about: the motivation to fail at a challenge, learn from it, and ultimately overcome it. The feeling of finally defeating a difficult boss after dozens of attempts is not an obstacle to the game, but its very core.
AI assistants can still be useful, especially in competitive games. Their potential does not lie in taking over tasks, but in acting as a kind of coach. Highlighting mistakes, explaining decisions, and suggesting ways to improve would empower players rather than relieve them of responsibility.
For now, all of this remains speculative. Whether and how the patent will actually be implemented in Sony’s future games remains to be seen. That artificial intelligence will play a major role in the future of gaming, however, is hardly in doubt. What matters is the form that role takes. Games are allowed to challenge, overwhelm, and frustrate players, because that is precisely where their appeal lies. If AI is to have a future in game design, it should not serve as a shortcut to the goal, but as a tool that makes the path toward it easier to understand.
What do you think? Let us know in the comments!
