Last week I posted about SpamArrest -- a popular new way to screen out spam. If you sends an email to somebody for the first time, SpamArrest asks you to pass a Reverse Turing Test to prove you're really human. It sends you to a stretched and distorted graphic image of a word, and asks you to identify it (that picture above is an example ). Theoretically, this neatly seals out the spam 'bots, because they can't visually identify graphics -- that's a uniquely human ability. Q.E.D.

Unless, of course, you're blind -- and you're using a text-to-speech converter to surf the web. In that case, this test screws you, as a story in today's CNET points out:

It seems that they have jumped on a technological idea without thinking through the consequences for the whole population," said Janina Sajka, director of technology research and development for the American Foundation for the Blind in Washington, D.C. "These systems claim to test whether there's a human on the other end. But it's only technology that can challenge certain human abilities. So someone who doesn't have that particular ability is excluded from participation. That's really inappropriate."

Precisely the point. If you're blind and using a text-to-speech reader, you are, in essence, using a 'bot of your own to augment your natural abilities.

And that is, when you think about it, a kind of mind-blowing gloss on the whole idea of a Turing Test. After all, the original Turing Test tries to to differentiate between humans and machines. The problem is that it assumes that these are two totally separate species. These days, we're increasingly becoming a cyborg race: Humans whose everyday physical abilities are enhanced or supplanted by machine sight, machine hearing, and machine cognition.

Maybe we need to add a new category to the Turing Test! The new goal should be a test that can figure out whether the person on the other side of the screen is a regular human, a regular computer ... or a blend of both.