<bovril> it collects data about the surrounding environment, then discards it and drives into walls
- Set Pattern the computer actually makes no decisions; all enemies will make the same moves every time regardless of what the player does. Most of the enemies in (simple) shoot em ups and platform games fit this category.
- A.I. Roulette again, the computer is not making decisions per se; it is simply choosing a move at random. At best, this seems like the character is thinking about their actions. This type is often seen in turn-based Roleplaying Games.
- Analytical, or Responsive the computer chooses a move based on the situation. The ghosts in Pac-Man fall into this category, which in 1980 was considered impressive.
It is in this third group that Artificial Stupidity can be found. This is when the AI can select a move for its character(s), and consistently chooses ones that are completely stupid. This might be included on purpose as a balancing factor, or to invoke Mook Chivalry, but it's just as often a result of Idiot Programming; the programmers simply didn't program the AI not to make that move, and when the AI evaluates its choices, the poor move looks like the best one. (It's far more likely that The Computer Is a Cheating Bastard will be introduced to compensate for Artificial Stupidity rather than the other way around.)
Artificial Stupidity is particularly visible in Role-Playing Games, be they turn-based games like the majority of the Final Fantasy and Dragon Quest series, or strategy-based games like Final Fantasy Tactics and Disgaea, simply because it is in these types of games that the decision-making process is the most important, and therefore, the most visible. It can potentially exist in any game involving an analytical or responsive AI, though, and the more analytical the game, the easier it is to get an AI that's, well, stupid. For instance, even good chess games can suffer from a version of this, called the "horizon effect".
Differs from A.I. Roulette because AI Roulette chooses moves randomly. Artificial Stupidity puts some "thought" in its moves, making the most obvious stupidities less likely but creating more consistent general incompetence.
Suicidal Overconfidence is a specific case of this that's usually less about bad programming or making the game easier than about allowing the player to have something to do.
The Escort Mission is often a variety of this.
The opposite of Artificial Stupidity is Artificial Brilliance, where the AI makes surprisingly good decisions that convincingly appear intelligent. See The Guards Must Be Crazy for this trope as relates to stealth games, and A.I. Breaker for when bad AI allows game-breaking exploits.
Note that, for the sake of argument, this trope typically only covers situations that a player can be reasonably expected to enter over the course of normal gameplay. It's hardly fair to blame the programmers, after all, if you use a cheat device to get special weapons ahead of time and the AI has no idea what's going on.
This trope is not to be confused with Obfuscating Stupidity, though some games that computers can inherently play well will use Artificial Obfuscating Stupidity to balance the difficulty. For example, the computer can mash buttons as fast as they want, but they'll pretend to be slower at it to give the player a chance.
- Action Games
- Adventure Games
- Beat 'em Up
- Fighting Games
- First-Person Shooter
- Party Games
- Racing Games
- Simulation Games
- Sport Games
- Strategy Games
- Wide Open Sandbox
- Other Games
- Real Life
- Glitch Techs: In-Universe, Glitch intelligence is usually related to their base programming, so they tend to act as if they're still in their game. While this is why most Glitches are dangerous, since they run by video-game logic and don't comprehend the dangerous consequences of their actions, or see real-life creatures and objects as things from their game, it also means that they tend to still fall into the patterns that they followed in said games, like periodically exposing a weak point.