AI in Videogames

 Der Guardian berichtet ueber den gegenwaertigen Status von Artificial Intelligence (AI = Kuenstliche Intelligenz, KI) in Videospielen und wie sie verbessert werden kann. Ansichten von Videogames-Entwickler Peter Molyneux sowie anderen auf diesem Gebiet fuehrenden Entwicklern sind ebenfalls im Artikel enthalten. Ich wuensche Euch viel Spass beim Lesen ;^) [...]
(show me)(don't show me)

<<
Inside IT
The hard-thought race for intelligent gaming
Artificial intelligence is the holy grail for game designers, but just how smart are current methods and what's in the pipeline?
Alexander Gambotto-Burke · Thursday April 19, 2007 · The Guardian

Gaming has a lot in common with everyone's favourite heiress, at least in the public consciousness: it's pretty, but dumb. And now that Microsoft, Nintendo and Sony have released their latest games consoles, that statement becomes all the more pertinent - next-gen games look great, but they play like something that could have been made a decade ago. While visual fidelity has advanced exponentially over time, the technology that governs how games play, react and adapt - the artificial intelligence, or AI - remains relatively rudimentary.

A handful of developers are striving to change this. The British designer Peter Molyneux, recently awarded an OBE, has spent his career trying to inject sentience and reactivity into games - and with his upcoming title, Fable 2, he thinks he's made significant progress. "AI is certainly the undiscovered country of games design," he says. "Any game genre - from hardcore shooters to the most story-driven adventure game - would be truly revolutionised by AI driving plot, characters and scenarios."

Gaming evolution
This is because presently, every action a player makes in a game has to be anticipated by the developers; the software will break down or simply remain stagnant if asked to do something that hasn't been pre-scheduled. More sophisticated AI would allow games to come up with solutions to player decisions on the fly.

"That's where the great wins are, when you start improving AI," Molyneux says. "If I as a player can do stuff in the world that is outside what the designer expects, and the game or game characters react appropriately, that's incredibly powerful. Suddenly, you think, 'Well, this isn't something that's just waiting for me to press the B-button; it's evolving around what I'm doing.' That is very emotionally compelling and, if you get it right, it can often be quite spooky."

That said, Molyneux doesn't believe AI can be solely responsible for intense, dynamic emotional experiences; they need to be married with what he calls "smoke and mirrors".

"You have to define what games developers call AI," he says, "as opposed to academic AI. There's actually very little true, academic AI in games. If I go along to universities and talk to professors of AI, they sort of laugh at us and our crude attempts at real-world AI. But my promise has always been, 'Well, good AI is what you see, not how it works.' Whether that's a mixture of true AI and an illusion is neither here nor there, because it's really about what it brings to the game."

Steve Grand - creator of the Creatures AI experiment, AI researcher, android hobbyist and also a recipient of an OBE - is probably the only person in the world with the distinction of creating a successful game driven by true AI. Which is why he's not particularly enthused by the progress his field has made in the games industry.

"AI isn't so much unappreciated as nonexistent," he says. "Most of what counts as AI in the games industry is actually a bunch of 'IF/THEN' statements. If a computer character doesn't learn something for itself then the programmer must have told it what to do, and anything that does exactly what it's told and nothing else is not intelligent. This is changing, and neural networks and other learning systems are beginning to creep in. But games programmers tend to devalue the phrase 'artificial intelligence'.

"This is mostly because the importance of AI in computer games is now widely recognised, and hence any attempt to implement it - including Creatures - gets hyped up pretty quickly. As graphics have improved, the behaviour of characters has got more and more embarrassing. When characters looked cartoon-like, any vaguely lifelike behaviour was impressive, but now that characters have fluid movements, realistic textures and complex facial expressions, they tend to engage different circuits in the players' brains. The better the graphics become, the worse the behaviour looks. So the need for good AI is well-appreciated. The snag is that none of us knows how to make it work yet."

Learning curves
Which isn't to say that Grand hasn't tried. In fact, Creatures is arguably the only AI-driven simulation that could still be classified as a compelling game in its own right. Basically a facade of evolution, the game allows players to teach their Norns (the creatures) to eat, talk, and defend themselves; they would learn and develop dynamically, and would pass on certain genetic traits to their offspring.

Grand tries to explain its technology: "It consists of a very simplified set of the building blocks of life - nerve cells, enzymes, genes and receptors. The hard part is explaining how all these virtual objects are assembled together to make something that lives, acts, breeds and evolves. But if you understood school biology - I mean, if you really understood it, not just learned it - then you already almost understand Norns.

"I think the basic approach - biological and bottom-up - is the only way we'll ever achieve truly intelligent artificial systems. But for the time being, it's not the best way to approach the problem for most games. If you want to reach for the moon you have to go to the trouble to build yourself a space rocket, but if you only want to lift yourself three feet off the ground you might as well jump. There are easier ways to make the kind of intelligence most present-day games require, even though these techniques will never lead anywhere."

Molyneux agrees. He's a fan of Creatures and the technology behind it, but he doesn't see it as particularly relevant to most mainstream games. "I have tried prototyping using neural networks," he concedes, "but you usually never quite get the results you want. And I think the system we use is a hybrid, really - I wouldn't want to put a title on the AI we've developed, but the theory of neural networks does influence it. I think the whole Creatures approach was incredibly interesting and impressive, but the real problem was that it was a bit of a strange beast. I didn't quite understand its motivations. It was a great, interesting attempt, though."

Even if Molyneux is sceptical about Creatures' mass-market relevance, his latest project, Fable 2, at the now Microsoft-owned Lionhead Studios appears to be heavily influenced by Grand's games. This is most obvious in the player's constant companion, his dog. According to Molyneux, each player's dog will be unique thanks to their actions. "We're really pushing the idea of a learning creature with this game," he enthuses.

But whether Molyneux will be answering that question with AI or "smoke and mirrors" is hard to say at this point; Fable 2 won't be released until 2008. Still, projects like this are an opportunity to advance our understanding of how AI can revolutionise increasingly stale gameplay.

Grand, who's now focused on robotics, remains confident that games can benefit AI research: "I think it's probably the best environment for AI that exists, at least until we've cracked some of the huge problems that are holding back robotics. When you write a game, your only responsibility is to be entertaining. It's not a mission-critical environment, so this gives you plenty of scope for new ideas."

With luck, more developers like Molyneux will begin taking his words to heart. It's unlikely entertainment software will ever employ - or need to employ - the sophisticated AI that Grand works with on a daily basis, but his work should serve as an inspiration to designers and programmers everywhere. After all, no matter how pretty games become, they're always going to be dumb if the brains aren't paid some serious attention.
>> # top # | Q: Technology Guardian.co.uk