Tag Archives: Marvin

Alan and Marvin – When machines talk

I’m a bit behind with my commentary on current news. But that’s okay, I think. Some days or weeks ago, the news was that Gordon Brown had issued an official apology to Alan Turing, a genius computer scientist, who was heavily discriminated against and “treated” for being gay until he committed suicide in 1954. Previously, Turing had helped crack the Germans’ encryption code during World War 2.

The story of how Brown’s apology came about is in itself noteworthy, for it came out of an e-petition on the UK government’s website. I must admit, I didn’t even know such thing existed in this country and I will certainly have a look at how e-petitions work. In this case, more than 30,000 signatures were collected and Alan Turing rightfully received his posthumous apology.

Turing Test of human “intelligence”

I guess Alan Turing was mostly known for his Turing Test. The Turing Test is an assessment of a computer’s ability to mimic human intelligence to such perfection that a human judge can no longer tell the difference between the computer and another human. It’s important to mention that the computer and his human counterpart are placed in separate rooms and they both talk to the human judge via some sort of text-based interface so that voice or handwriting don’t influence the verdict. The human judge can ask all sorts of questions through that interface to figure out who’s human and who’s not.

Every so often, some bright minds actually have a real competition based on the Turing Test. The contestant with the best human-intelligence-mimicking computer wins. Don’t know what though. Probably money and three levels up on the geek scale.

Of course, you may ask, “Wait a minute… you call being able to communicate through some text-based interface ‘intelligence’?” The philosophical debate about this is as old as the Turing Test itself and if you expect an authoritative answer in the next few paragraphs, I’m afraid I must disappoint you.

My humble opinion would be that simply being able to carry a nice conversation through some chat program doesn’t mean the computer in the other room is “intelligent”. Intelligence means relating the words and sentences of the conversation to the situation and the context in which it takes place, to everything that happened before and everything that is likely to happen later, and finally, relating the words and sentences to a “being” (don’t make me define “being”…). The conversation only makes sense when you interpret the words and sentences in relation to the unique “being” who expressed them.

The author JD Peters put forward the argument that Turing reduced intelligence to communication without the presence or interference of human bodies because he (subconsciously) wanted to escape the stigma of being homosexual. He may have hoped for a form of interaction and communication that is not distorted by any human “flaws”.

Speaking of talking machines…

While writing this, I remembered that the science fiction literature has already created a number of intelligent machines or robots that would not only ace the Turing Test but also live up to the standards of intelligence that I tried to describe just now (R2D2, etc.).

My most famous example is Marvin, the paranoid android from the Hitchhiker’s Guide to the Galaxy. He’s an incredibly intelligent robot, but he’s also incredibly depressed. Here are just three little quotes…

Marvin: “I am at a rough estimate thirty billion times more intelligent than you. Let me give you an example. Think of a number, any number.”
Zem: “Er, five.”
Marvin: “Wrong. You see?”

“I’d give you advice, but you wouldn’t listen. No one ever does.”

“I think you ought to know I’m feeling very depressed,” Marvin said.