Blog # 110…October, 2020

I’ve been thinking a lot about artificial intelligence lately, picturing it like a sly little demon lurking in the shadows ready to pounce and take over some part of my life.  Notice that it’s not alternative or replacement intelligence but artificial…forced, contrived, sham, stilted, all imply a copy of something natural, not as good as the original, think artificial flowers or vanilla.

The word artificial has its root in art - created by humans - reflecting natural forms like the rendering of animals on cave walls, progressing to landscapes, portraits and still life. One definition mentions lack of spontaneity.

So, I was intrigued recently to hear Janelle  Shane, who has a PhD in electrical engineering,  a MSc in physics and a delicious sense of humour,  speak about her new book,  You Look Like a Thing and I Love You.  Although it's not a great pick up line,  I did get the book and was amazed (and comforted) by what I learned but most of all it gave me a great number of uproarious laughs…a rare commodity these days!

So, you may be interested in the comforting things – AI can only perform in a narrow, tight path, can’t do critical thinking and can’t do much at all without humans...whew!  Janelle quotes leading machine learning researcher Andrew Ng, “Worrying about an AI takeover is like worrying about crowding on Mars.” Not in our lifetime or probably the next either.

It was also comforting to see the difficulty AI has with learning simple things we find easy, like composing Knock knock jokes or creating recipes. That’s where the belly laughs came in - the recipes it came up with in particular, with titles like Basic Clam Frosting or ingredients like peeled rosemary and six tablespoons electric mixer or Spread Chicken Rice where you could use either 1cup of shredded fresh mint or raspberry pie…no mention of chicken.

There’s a lot about self driving cars which are much in the news and have a long way to go before they can be without an alert human to take over when the algorithm doesn’t recognize an object or situation. They're closer on the horizon and bear watching.

The takeaway here is that what we have now is artificial narrow intelligence, what we often see in books and movies is artificial general intelligence. So if you see a robot that can beat you at chess, tell you a story, bake you a cake and name three things larger than a lobster, it’s solidly the stuff of science fiction, nothing wrong with SF, just don’t believe it’s here now.

Last month I threw out the idea of sharing stories of the past 6 months and got four people ready to go, so we'll be starting a thread in the next few weeks. Lots of room if you want to join, to write or to read, just let me know and I'll weave your address into the cloth of experiences in this covid world.

Back in the cruellest month, or can it be the kindest? Let's try for that.  

And let's see The New Yorker's take on AI, hmmm maybe we've known people like that too.


But wait, there's more. This just in from the University of Alberta...in case you're still thinking about aboriginal issues and what to do about the Truth and Reconciliation Commission report, Uof A is offering a terrific course through Coursera called Indigenous Canada. I started it this morning, and apparently so did Dan Levy, another reason to love the guy!







No comments:

Post a Comment