News Flash: Games now have "artificial intelligence"
Ok, we all know that journalists aren't subject matter experts. That's fine. However, sometimes they are blatantly uneducated about the stuff they cover... and if you are in whatever biz or industry they are talking about, errors jump out at you. For example, I know quite a bit about the aviation and airline industry. You would not believe the pathetic errors that journalists make about those areas when often the answer is as far away as Wikipedia.
Anyway, my Google Alerts picked up an E3 review from a newspaper called the "Straits Times". It seems to be based in Singapore for what that's worth. Anyway, check out the opening few lines of the column Videogames getting minds of their own (emphasis mine)...
LOS ANGELES - VIDEOGAMES are getting smarter with virtual enemies improvising during battles, storylines shifting based on moral choices and in-game characters sending players text messages for help. Titles unveiled at the just-concluded Electronic Entertainment Expo in Los Angeles feature artificial intelligence (AI), making in-game worlds more realistic and less predictable.
Wow... games now feature artificial intelligence. That is soooo cool!
At the recent Paris Game AI Workshop put on by AIGameDev (et al), Phil Carlisle spoke about automated testing in game development. His major point was that, by decoupling the game engine from the display engine, you can actually run the core simulation very rapidly. It also takes far less processing power. You can run hundreds, thousands, or even millions of simulations during overnight hours, for example. By logging the resultant data, it exposes glitches and imbalances that you may have overlooked.
Another important part of this method of development is that it allows the engine developers to work on things long before it is ready to go on-screen. Some types of games can be "viewed" if not actually "played" on a regular DOS window. This is what I did in my initial development of Airline Traffic Manager. I was able to do a lot of work on the core algorithms such as the passenger generation, moving the aircraft across the map (in Great Circle routes!), the basic state machine interactions between things such as the aircraft and the gates, etc. When it came time to hook up the display, the game engine was ready to roll.
ThinkArtificial on Game AI and Apparent Intelligence
It's always interesting when game AI gets commented on by non-game AI folks... especially the academics. On a relatively new blog/site, ThinkArtificial, the author commented on how games are, in general, "horribly void of intelligence." I suppose that's often the case - particularly in the 2nd tier of the market. However, I wonder if he went over the top a bit. (Note that my friend and colleague, Paul Tozour, jumped right in the next day with the first comment.) Anyway, here's his opening paragraphs.
A subscriber of Think Artificial wrote to ask me about games and AI. In short, DF asked what my thougths are on AI in games and which ones I think are the most intelligent.
To answer this bluntly: Game AI is very different from it's non-game counterpart, and it's not my field of study. I've only compared modern games through a window. However, Alex of AIGameDev has superb coverage of AI in games and the top AI games of 2007, by community vote. The top of the line are Half-Life-2.ep.2 and BioShock.
But regarding Game AI in general: modern games are horribly void of intelligence. It depends on where you set the bar, certainly. There's tons of AI in modern games compared to 5 years ago. But the first thing to note is that Game AI is not the same as AI. It's a subset of it. Just like discrete mathematics are a subset of mathematics. And moreover, Game AI is a very specialized subset—it has well defined goals, models for construction and limitations.
Thankfully, he did acknowledge in the article that game AI is somewhat hampered by the budget and the needs to make for fun gameplay rather than excellence.
Anyway, at least it wasn't derisive like we sometimes get from academia.
If you are a devotee of the Civilization franchise or if you are interested in where the upcoming game Spore is headed, there were recently two interviews with Soren Johnson.
The first was at 1up.com. The second was at Gamasutra. In both, he talks about how his experiences with (and distinct interests in) strategy and simulation AI translated over to the world of Spore.
I happened to see this trailer at GameVideos.com. It is from the upcoming game "Legendary" from Spark Unlimited and Gamecock Media Group. While you can't garner too much from the video about the intelligence of the enemies, I did find it kinda groovy. What's more, I think it is interesting that game studios are releasing video trailers that are specifically designed to showcase AI. Before, it was about the graphics, or the physics, or the storyline... but never the AI. Maybe we are on the cusp of an explosion in the AI world.
This isn't necessarily directly AI-related... but when I read it, I thought about how appropriate the process is towards making design decisions about AI characters. On the Team Fortress 2 blog, they have posted about the process they are going through in the creation of a class pack for the Heavy. In it, they list some of the considerations that the team makes in determining what should go in and what should stay out. Here are some examples:
Goal: Make the Heavy more viable when he has no Medic to pair with.
Constraints:
* It shouldn't have a cumulative effect when being healed by a Medic as well. Heavy/Medic pairs do pretty well as it is.
* It shouldn't significantly change the Heavy's role, relative to other classes. In particular, it shouldn't significantly encroach on another class's role.
[snip...]
Extra:
[snip...]
* Is it an interesting tool to choose relative to the base Heavy weapon it's replacing? What scenarios can you envision in which each is useful? What arguments can you raise for why each is better than the other?
* How often does the Heavy need to think about it? Is it something he uses once every 5 minutes, or is it something he needs to be constantly thinking about? A greater impact on player decision making is generally a good thing.
[snip...]
Finally, keep in mind the skillset required to be a good Heavy. He doesn't really rely on fine aim, since his minigun has such a wide spread, instead relying on more tactical skills, like these:
Being in the right place before he starts firing, because he has little ability to move while firing.
* Good anticipation of enemy behavior, for both the above point and because his minigun needs to be spun up before firing.
* The ability to estimate the amount of damage he's taking. It takes time for the minigun to spin down, so he needs to be able to know when it's time to retreat several seconds before his health has run out.
[snip...]
There are a few that I cut out of there. Go to the link above to see more of it and some description of the whole process.
The reason I find this interesting is that they were very cognizant of taking a holistic approach to the game... and to the game players as well. i.e. who is going to choose to play a Heavy, and why? Can we magnify that? What will it break? Does it step on the toes of other classes?
Anyway, excellent questions and a nice peek at what it takes to be a game designer and developer.
Got linked to this by Paul Tozour. Here are Damian Isla's (Bungie) slides from his 2005 AIIDE presentation on "spatial competence" entitled "Dude, where's my Warthog?" It includes info on a ton of the stuff he/they did in Halo 2. Included is information on pathfinding - especially with regard to how we (as people) process spatial information. It's nice to see someone else tapping into psychology as a source for potential solutions for game AI.
Fantastic stuff... and left me dying for the audio.
