My friend Michele just posted this hilarious item on her blog, Hippo Dignity. It's a screenshot from today's Google News. It appears that Google News concatenated an item about how Cherie Blair is writing a book -- with various articles discussing how Tony Blair is scheduled to appear on The Simpsons. Oops.

I actually am an enormous fan of Google News -- not least because it does a superb job of collecting news from a plurality of sources worldwide. (I'm as likely to find stories from Australia, India, or smaller U.S. papers, which helps break my monotonic reliance on The New York Times.) And actually, I'm also strangely fond of these occasional glimpses into the weirdness of machine logic. They're like Freudian slips. After all, we humans, too, make all manner of intriguing and revealing verbal mistakes, and they can seem funny or odd. But once you unpack these mistakes, you realize the deep psychological reasons we make them, and you understand we really meant to say.

So we could consider this Google-News mistake a mechanical version of a Freudian slip. The computer says something so incredibly weird that it startles you, and only upon closer investigation do you learn more about why the machine did it; and then everything makes a bit more sense. The machine's facade of Turing-like intelligence slips for a moment, and you spy the algorithms at work. I think that's a good thing, because it's socially and politically important -- to say nothing of really fun -- to understand what makes our intelligent machines tick.

Actually, here's an idea. Since there are so many of these little machine mistakes every day, we should have a term for them. We need a name for the artificial-intelligence equivalent of a Freudian slip. So I'm inventing a new term, right now:

The "Turing Slip".

You read it here first!