In 1950 mathematician and computer pioneer Alan Turing proposed a measure of computer intelligence that came to be known as the Turing Test.

Turing wanted to replace the philosophical question of “Can computers think?” with a more practical, and testable, measure. So, he devised a test where a human would type messages back and forth to an unseen correspondent. If the human couldn’t tell whether or not the “person” on the other end was human or computer, then effectively we should consider the machine intelligent. After all, isn’t our assessment of the quality of the communication with another entity the essence of our measure of intelligence?

While many have debated the usefulness of the Turing Test, no one can argue with power of its philosophical thrust. The Turing Test simultaneously raises issues around the definition of intelligence, the nature of consciousness and intentionality of thought as well as the phenomenological relationship between people and machines. That is to say, once computers can easily and regularly pass the Turing Test, what then? Whether we think of the machines as intelligent or not, we will live in a world where human and machine intelligence have become indistinguishable in certain circumstances.

What makes this fascinating to me at the moment is a simple thought I’ve had which leads in a hurry to a surreal and challenging intellectual landscape:

All video games are a form of the Turing Test.

When you sit down to play any video game—whether on a console or computer—you interact with a system that responds to your actions and you respond in kind. You assume that you are playing against (or with) “the computer”. But what evidence do you have?

This question seems simple and obvious on the surface. Because, if the machine isn’t connected to a network, and no one else is sitting there playing with you, then it must be the machine. Right?

While is a bull-headed empirical attack on the problem, it doesn’t address the issue at the heart of the Turing Test. In Turing’s test, the only evidence you could use to reach your conclusion was the information typed into the test and the typed responses that were returned. He didn’t think it would be a meaningful measure of intelligence for someone to observe that that the machine didn’t look like a person so wasn’t intelligent. He wanted to control for variables that didn’t directly relate to his measure of intelligence—the interaction in conversation.

Applying this information-constrained model to the video game example, you can’t simply look to see if the Xbox is connected to the network or crack the case to see if some ingenious robot intelligence snuck a wifi component into your machine. No, you have to figure out if the thing you are interacting with is a computer or a person based on the interaction.

The main difference in the video game as a Turing Test and the test Turing designed is that Turing wanted to use a conversation. In the video game, the conversation has been expanded to include all interaction with the system. (See Chris Crawford On Game Design for a particularly apt description of interaction as a form of conversation!)

Now, if this version of the human-or-not Turing Test seems a little hard to swallow, then simply clarify the situation a bit—You’re now playing a game of Madden on a PlayStation against an unseen opponent. Computer or human? How can you tell? Or, you are playing a game of Everquest and have found yourself dealing with a particularly impish member of an elf tribe who seems intent on getting on your nerves. Human or computer?

Now back to, some other game examples:

Tetris. Is it a human or computer setting the order of the piece that will fall?

Chess. Sure you installed the program on your hard drive. But how do you know it is a program and not some remote chess master matching you move for move?

Halo. Is the Covenant under computer control or being manned by a staff of temporary office workers in Bangalore?

Resident Evil. After you play it a couple of time, you get the idea that the game is running on script. But the first time through, is there anyway to tell that there is not some malevolent director pacing things as carefully as possible to shock you?

Without checking the known physical routes for interconnecting human players, how do you know whether or not the agent you interacting with is human or computer?

These examples may seem terribly trivial, and even baiting the question of intelligence. Certainly, some critics of Turing’s initial test have argued along those lines. But as I said initially, once you cannot verify that you are interacting with a human, you have fallen into an epistemological trap with the most obvious exists being market—“Computers are intelligent” and “Our notion of intelligence is wrong”.

On its own, I think the point that video games are a type of Turing Test is helpful in connecting the philosophical issues surrounding the Test to the subject of video games. And along those lines, I’ve discovered a possible way to connect information and complexity theory into critical video game studies.

In a nutshell, we find more complex games more satisfying. One way to ensure complexity in a game is to make human input integral to the play of the game. Human behavior, almost by definition, is complex. So, a fairly simple game rule-wise, like chess, becomes interesting when played against humans because they bring in a degree of complexity not present otherwise. Along these lines, randomness is not a substitute for complexity, so rolling of dice tends not to create interesting complexity.

If we follow our conclusion about the Turing Test, then playing a game with sufficient complexity provides an equivalent experience to playing with or against people—at least on the non-social, rule level. This seems to me to be central key to the attractiveness of video games. We may not think of something as cold an inert as a computer as alive, but we may actually enjoy them because they behave intelligently! Or, to put it in slightly different terms, interesting interaction is always a conversation with intelligence.

More on these subjects to come!