A Rant about A.I

A.I in the context of something-or-other stands for "artificial intelligence", and Wilkipedia (inter alia) will explain it and link you to scientific and philosophical discourses on the subject.  Well, DREAM ON! As I type this, the "Radio Times" is on the coffee table in front of me. But it is "upside down". Nevertheless I can still read the title.  As I could if the idiot printers had printed "Radio" upside down with respect to "Times" and the whole thing reading from right to left.  Or with every letter printed in a different colour and in a different typeface and with some of the letters upside down and some the right way up and and some of the letters mirror images of themselves and with the odd letter printed twice or in the wrong order and maybe with all the vowels missed out.  (Though not so easy if all the consonants were missed out.  Hmm? That's food for thought, innit?)  Can a machine reader do any of these "human" intelligence tricks? The root of the problem is simple.  The nerds, geeks and A.I Departments in our tenth-rate Techyversities begin by defining "intelligence" as a "something" which they can make a machine simulate. Like - recognising a chair as something to sit on, or walking downstairs without falling over. Then they say,  "Look at my wonderful machine.  And more imnportant, LOOK AT ME!!  Aren't we wonderful?  Aren't we intelligent?"  Well . .  er . .  no!  The techytype with his soldering iron and "Unix for Dummies" textbook may well be, but his quasi-useless machine isn't.

Here's another example of an artificially intelligent system whose I.Q is somewhat lower than a stone.  The other day, I was required to "activate" a new credit card. On keying the 0800 number (freephone to my friends across the Pond), a nicely spoken person asked me what I wanted to do.  Usually they tell you which button on the phone to press, and for a glorious moment I thought I was speaking to a real live girl. But NO!  When I said "I want to activate my new credit card", the voice-simulating machine replied "I'm sorry.  I didn't understand that.  Please try again."  Very slowly and carefully I said,  "Activate. New. Credit. Card"  This went on through several repeat loops.  (Oh . . . I'm sure I wasn't enunciating English any worse than "she" was, and I certainly speak more clearly than most of the interviewees and interviewers that you hear these days on radio and TV resolutely honking and snorting their way through one of the world's most beautiful languages. If you want to hear good English these days on the BBC or BBC World Service, listen out for the Swedish truck drivers and Chinese physicists. They speak English better than the average Estuary Native.)  But I digress.  Eventually "she" gave up and asked me for my card number.  Guess what?  Same routine. No matter how a enunciated the sixteen digits, "she" told me she didn't understand.  Finally I was asked if I would like to "speak to one of our experts"? . .  and wey-hay!! - "she" managed to understand my "Yes"  The point is, of course, that if I had been connected to a normally intelligent human in the first place and had said (something like) - "Oh, hi!  How yer doin'?  Listen, see, er, er, I got my, er, what's it, um, new credit thingy, credit card, you know, and I need to, like, hang on where's it gone, right, got it. What's the word?  Yeah. Activate it, you know, the new card thingy business."  If I had come out with that gobful of ordinary human speech, the "intelligent" human at the other end would have decoded it instantly into what I wanted to do. Artificial Intelligence? Perleeease.  Don't make us laugh!