Player empowerment sits at the heart of EVE Frontier. Indeed, this is exactly the frame its game director FC Goodfella uses to explain how the studio is approaching making a blockchain game for a world of abundant artificial intelligence.
Whereas most developers talking about AI first discuss productivity or content generation, the dev update “AI, Automation and Agency on the Frontier” treats AI mainly as something players and the world will use to play, hence as a design problem to be solved openly rather than policed behind the scenes.
The premise is that EVE Frontier is a game about survival through tools.
Players can reshape the world in real time using Smart Assemblies. The update positions AI as an extension of that toolkit: a way to build complex mods with less technical knowledge. For example, during the recent Smart Assemblies hackathon, players used AI alongside the game’s systems to turn ideas into working mods faster, some producing work that would previously have demanded considerably more technical skill and time.
But anyone who has spent time in the EVE universe knows that powerful tools get used in ways their designers never anticipated. Botting, exploits and increasingly sophisticated automation are the current form. Trying to fight these purely through detection and enforcement becomes an endless arms race says Goodfella, something that drains engineering time away from making a good game.
His alternative is to design for automation rather than against it, accounting AI as part of the ecosystem. The example given is multi-boxing. Rather than outlawing it, the studio has designed moment-to-moment piloting and combat which is so demanding that a player’s attention cannot drift away from playing a single instance.
In this manner, the friction lives in how hard the game plays, not as an arbitrary rule to be enforced or exploited.
Underneath sits the fundamental principle that intelligence should exist “within the Frontier, not above it.” Everything in EVE Frontier costs energy and resources, and that holds whether the decision behind an action came from a human or a machine. For example, fuel forces real costs onto automation. Line-of-sight, occlusion and imperfect scanning data mean any actor must explore, scan and expose itself to risk to know what’s happening.
Under these conditions, the question becomes whether automation can perform well when handed the same risks and boundaries as players, and what happens when it chooses wrongly?
The update then turns to discuss AI as creative opportunity. Games have historically faked intelligence with scripts and predefined behavior, it notes, but that is changing fast.
The in-game feral AI already haunts the Frontier, left over from the catastrophe that ruined human civilization. Max Cats imagines versions that scout, organize at scale, react to surrounding events, retreat from players they judge too dangerous, regroup and pursue objectives unrelated to you unless you interfere. Beyond combat, he proposes scouts sent into unknown space, drones doing specialized work inside a base, a ship assistant, and contextual help for new Riders that guides without taking away their decisions or their mistakes.
Crucially, each entity remains bounded by the same rules. A drone needs energy, a scout needs reliable information, a ship assistant should not magically know everything happening across the universe. All of it should be constrained by the same Digital Physics as everything else in the game.
Of course, with EVE Frontier still in development, this remains more a manifesto than a blueprint of how things will work. Yet, as a statement of direction, it is unusually specific. Developers should now treat automation as a fact and place it within the same physics that govern players, making intelligence interesting by refusing to let it sit external to the rules.