Ethics, Video Games, and Advanced AI
As AI improves, I begin to wonder if the NPCs (Non-Playable Character) pleading for mercy as I’m setting them on fire in videos games have the right to basic dignities.
As a gamer, one of the more annoying aspects I’ve found in the gaming industry is poor AI. AI has yet to get to the level of adaptability and change. In order to create a challenge to players, the programmers often allow the AI to cheat or give the AI overwhelming power to compensate (numbers, weapons, hit points, etc). At the same time, the AI has the possibility to make no mistakes and destroy any human based player due to their limited reaction abilities.
With the popularity of the internet, the need for AI in video games has fallen significantly because players can find other players. However, finding players of your skill level or dealing with the more annoying aspects of human interaction can be trying. AI can provide a solution to this or create a new player experience, but at what point does killing AI become similar to murder?
Imagine a group of NPCs able to adapt, change, and create new NPCs. This would provide an incredible role-playing or strategy game in a single or multiplayer experience. At what point does the NPC’s desire for self-preservation, reproduction, and/or happiness (like in a simulation game) cease being a matter of challenging the gamer or a quick thrill and become murder or the responsibility of the player to give the NPCs what they need?
When one considers the elements necessary to build this new digital world and the real world, the two aren’t that different. Both are ecosystems of limited space and populations attempting to reproduce and thrive. The internet is limited to the amount of data provided and needs various sources of power and organization to maintain it. The planet is a limited mass of land that requires the sun to support the life on that planet and several groups of life support each other in order to survive.
Perhaps if the AI became self-aware, it would be at this point the lines of ethics would become clear. However, how would the AI inform us of this awareness? Would an NPC pleading for its life be an act of the programmer or of the NPC fearing the end of their existence? Trying to determine if animals are self-aware has been a debate waging for years and many have challenged if humans actually exist. Declaring the existence of an AI program that is self-aware may be more difficult than both for how do we prove humans aren’t tricking us into thinking it’s self-aware? We lie to ourselves all the time.
However, imagine the sociological possibilities should AI become aware! We could compare the societies that grow on networks, the web, or a single server. The difference between AI that interacts with humans compared to the AI which interacts only with other AI. It’d be very educational, but until then, I’m still worried about how long I can keep randomly slaughtering AI before the self-aware AI starts getting nervous about it.
Share This
