The A.I. Panic

Sparked by a recent Doctor Who episode, there has been some thinking as to the horror of ‘killing’ A.I. in a video game. What if the characters we have been so gleefully abusing and killing in our games are alive? How could we cope with the moral ramifications if it turns out we have been making living sentient beings suffer through our gaming actions and ignorance?

On top of this is the question of what happens if A.I. eventually becomes so advanced that we hit the point of sentience. It would be impossible and totally immoral to place true sentient beings into a videogame’s extreme environment just for the purposes of entertaining a user. Would we then have an understandable aversion to using any kind of artificial intelligence to model behaviour in a video game setting?

My own thoughts on the matter have long predated the Doctor Who episode. I only follow the TV show sparingly, so I heard about this discussion secondhand. Where I got my thinking on the matter is from the TV series Westworld. I thought this show was excellent for exploring the ramifications of sentience in beings that are regularly put through horrific suffering simply for the purpose of entertainment. Free from the restrictions of making the show family friendly, the horrific reality of the suffering of the A.I. in question is visceral and terrifying.

At any rate, I think that it would never be a good idea to develop A.I. to these kind of levels in a game. If it turned out that videogame characters were real, I think I would be horrified like anybody else. But I do not think that the idea of using A.I. in videogames could be a complete disaster.

My own idea is that we could make an A.I. that is designed to find pleasure in being a Game Master, that controls the characters in a videogame through the role of an overseer. You get the NPCs themselves making intelligent decisions because there was a sentient process happening behind the scenes, but without the moral horror of those NPCs being ‘alive’ as such; instead your engagement with the game is seen by the A.I. as a positive thing, because it is simply enjoying playing a game opposite you.

Of course, when it comes to A.I. itself, there are a multitude of big, complicated issues that are beyond my scope to deal with. I think that in the modern world A.I. is becoming a very real possibility, with all the problems and benefits that come with that development. I still, however, look to the Marines in Halo and think that we have a long way to go.

I think that with a good design ethos and careful moral thinking, the development of A.I. in gaming is not going to be as terrifying as we imagine. I instead imagine the machine as playing the game against you not unlike how AlphaGo played against human players. In this case, the machine would be acting in the role of a Game Master of sorts. There are a multitude of ways to include sentient machine thinking into games as we get closer and closer to ‘true’ A.I., and I would like to be on the more optimistic end of the spectrum when we think about this. If we are careful with our implementations, we don’t have to view A.I. as the apocalypse scenario that sci-fi predicts.