“011010100,” Said Voight and Kampff: When Robots Administer Turing Tests

Heliopoli_Full_EX96In the “It’s probably obvious to everyone except the chief archivist — until now” department, the excavators would like to point out that Turing tests do now exist, we use them often and it is the machines who are testing us.

A brief interlude from the movie Blade Runner:

Rachel: Have you ever retired a human by mistake?

Deckard: No.

Rachel: But in your position that is a risk.

Tyrell: Is this to be an empathy test? Capillary dilation of the so-called blush response, fluctuation of the pupil, involuntary dilation of the iris …

Deckard: We call it Voight-Kampff for short.

Tyrell: Demonstrate it. I want to see it work…. I want to see it work on a person. I want to see a negative before I provide you with a positive….

Deckard: On you?

Tyrell: Try her.

Back to programming.

In 1950, Alan Turing, considered the father of modern computer science, proposed a test of a machine’s capability to demonstrate intelligence. The Turing test is whether you can tell that the entity on the other end of a computer link is a┬ámachine or not, when engaged in a dialogue only by typing text. If a machine is on the other end and you can’t tell the difference, the machine passes it’s Turing test for artificial intelligence.

Often, now, when registering with or logging on to a company through the Internet, we must enter into a text field a word that is displayed in a graphic, all skewed and squiggly, because the spam bots and spy bots or whatever can’t see it. It’s a test to determine if a human is on the other end, looking at the screen.

Heliopoli_Full_EX97Some “send a message” boxes in Web sites and blogs now have an extra field to fill in, to prove that you are not a bot that wants to send comment spam. You have to answer a question such as “What is 3 + 3?” (Of course, adding 3 + 3 is the simplest thing a computer can do, but that’s beside the point.) We’ve even seen one that asks, “Are you human?” That one certainly gets to the heart of the matter.

These are Turing tests. The questions in the Turing test (“We call it Voight-Kampff for short”) are: Do you see the hidden word? Do you know the answer to the equation? Are you a human?

But what happens when the interface, the intercessor, is a machine? We’re not talking about a laboratory situation, in which one is typing into a computer and getting a response through a linked computer behind a screen. We’re talking about the Internet, which has additional computers as go-betweens. The administrator of the test isn’t a human, is it?

The excavators — and the chief archivist, now that he’s hip — think that this is all ironic somehow, and funny and cool and absurd. And eventually the Turing test will have to get smarter as the spam bots get smarter: “Is this to be an empathy test?”

The computers are testing us, to see if we are human. And what happens as those intercessors get smarter?

Heliopoli_Full_EX95For now, the machine, through another machine, reports back to its owner, “Yeah, he answered ‘6,’ he passed.” But couldn’t the machine, having been influenced by a smart or corrupted intercessor computer, just as easily report back, “Um, well, he said ‘7’, yeah, that’s the ticket, he said ‘7,’ I tell ya, and he ain’t no human neither. Harumph.”

We’re just saying.

The excavators are thankful that they do not blog, but merely uplink to the chief archivist, who sorts through the transmissions and determines what gets posted, and who couldn’t possibly, ever, we’re sure of it (we are!), be a machine.