Science Journalism is Misleading at Best

So, I just read this article. It's a typically blurry piece of science journalism sadly.

Michael McGuigan thinks that computers will soon be able to pass the "Graphical Turing Test." The article makes him out as being of the opinion that real-time rendering is the main obstacle to passing the "Turing Test". In fairness to him, I doubt very much that that's what he thinks. Other scientists are quoted as saying things like this:

But others think that passing the Graphics Turing Test requires more than photorealistic graphics moving in real-time. Reality is not 'skin deep' says Paul Richmond at the University of Sheffield, UK. An artificial object can appear real, but unless it moves in a realistic way the eye won't be fooled. "The real challenge is providing a real-time simulation that includes realistic simulated behaviour," he says.

Which I'm sure no one doubts.

Imagine a virtual world where the user is playing a game of pool. This would be much easier to do realistically than a world where interacting with a virtual human (or indeed even a dog or cat, for example) was possible. In order for the user to not realize that she's in a virtual world, that character would have to behave naturally, of course. Thus, we're back to the "Turing Test."

So, I don't know if this author is conflating "Graphical Turing Test" with "Turing Test" or what's going on. The only thing of interest that this article says is this:

"You never know for sure until you can actually do it," he says. "But a back-of-the-envelope calculation would suggest it should be possible in the next few years, once supercomputers enter the petaflop range – that's 1000 teraflops."

That's McGuigan talking about being able to pass the "Graphical Turing Test." So, a nice VR is coming soon. Not a machine capable of passing the "Turing Test."

This is a sort of sensationalist journalism. It's just aimed at a geek audience. Sensationalist journalism is sensationalist at best, and dangerous at worst. Please read Manufacturing Consent if you haven't already.


Note: I was an intern at BNL during the summer of 2003. Michael McGuigan was my mentor while I was there.