In yesterday’s blog post, I wrote about gardening being A Strange Avocation, and I explained the difference in the behavior of flowers between this year and previous years. The tags I used for the post are gardening, home, life, bees and flowers, bees in a garden, larkspur, lily trees, purple echinacea, purple echinacea turned pink, yarrow. All garden related and directly related to the article I posted.
WordPress, the platform that hosts this blog, has an AI tool that suggests various tags for posts. In case you don’t know, tags are words or phrases that bloggers use to help people find their articles. For example, broad tags such as “gardening,” don’t do much to drive people to a blog post since they are too general. If someone were to Google “gardening,” they’d get millions upon millions of hits, and the chance of their finding any particular article are close to nil. Specific tags are best, or so they say, because the chances of someone stumbling upon your immortal words are a lot better. (I was being facetious about the “immortal words,” making fun of the fact that so few of our blog posts say anything of importance, and yet, in the electronic age, everyone’s words, no matter how puerile, truly are immortal.)
You will never in a million years guess the words that the all-seeing, all-knowing AI suggested for that gardening post, so I’ll just list them for you: “Auto immune disease, breastfeeding, ulcerative colitis, Crohn’s, IBD.” How is that even possible? I’ve gone over and over what I wrote yesterday, and I cannot see anything that could possibly have generated such a wildly inappropriate group of words.
I know artificial intelligence is supposed to be a great thing, and it might be for other people, but I’m not that impressed. From what I have seen, generative AI in particular is not any more intelligent than most humans seem to be. (I slipped and called it degenerative AI before I corrected myself, but I like that term!) It also has huge biases, especially political ones, which makes sense, I suppose, since the vast majority of news sources and sites like Wikipedia are liberal, and generative AI is only as smart as its input. It would be nice to be able to use Google’s AI generated summaries for a quick perusal of any given topic, but I can’t rely on its answers because it makes mistakes that I know are mistakes. It doesn’t always understand the question, either. I wanted to check out “immortal words” to make sure I wasn’t making up the phrase, and all the AI would tell me is what immortal meant. I also wanted a quick definition for “tags,” and what I got was a whole lot of information about where to buy paper tags with strings. (I’ve noticed more often that search engines seem to be geared to products first, so if I want to know the definition of something, and it happens to be the name of a musical group or a brand name or whatever, that is what shows up, not the more generic term that I was looking for.)
I suppose, in Pollyanna-style, I should be glad that I didn’t need the suggested tags. Glad I don’t know enough about those medical issues from a personal angle to write about them. But still . . . utterly bizarre.
Actually, I’m not one to talk about intelligence, artificial or otherwise. Earlier, I was doing a puzzle, and I needed an eight-letter word beginning with L for a flower. And all I could think of was lavender. Lavender? Really? When I’d spent hours in the morning clearing away larkspur? When I’ve written post after post mentioning larkspur?
Oh, well, such is life. My life, anyway.
PS: The recommended tags for this post are Project Life, Heidi Swapp, San Diego Comic Con, Becky Higgins, Studio Calico. I don’t even know what most of that is!
***
Pat Bertram is the author of Grief: The Inside Story – A Guide to Surviving the Loss of a Loved One
Bertram's Blog 