by Beth Dolinar, contributing writer
When Wednesday comes, I put my column thinking cap on and wander around in my brain for a topic. For 25 years, I’ve been writing a column for a daily newspaper, a composite of about 600 words that explore an idea that I’d been living with for the past week.
It’s how I decide not only what to write, but how: how to make it readable and interesting—and sometimes even entertaining, funny, edifying or, rarely, profound. When people ask me where I get the ideas for my columns, I tell them that as I go through my life, I’m always writing the first lines, always feeling the words in my ears as I try them out. Which of those ideas have enough depth to sustain a theme?
The creators of Artificial Intelligence want us to believe that computers are reaching the point at which they can do that as well as I can. AI apps such as ChatGPT are designed to write copy. A complex system of algorithms allows the apps to gather information based on human specifications, and then arrange the data, and the words, into copy that makes sense. As the AI writing apps become more sophisticated, we’ll be seeing more examples that compare AI-generated copy to the copy that springs from a living brain.
I’d like to think I would be able to discern the real from the fake. If we’re talking about some basic informational copy, I might not know the difference, because informational writing is the most basic, and the first kind a writer will master. The AI apps search for the data, arrange it in sentences with correct grammar and spelling, and present it as “writing.”
But it’s not. Although AI “poetry” and “novels” are cropping up on digital reading platforms, they aren’t creating much buzz beyond the controversy about how they were written.
Stories don’t emerge from lines of code, of course. Stories spring from the places where humans roam: from our kitchens and our cars, from our bedrooms and our back yards. From the dog park and from the beach, and from neighborhood pubs and from classrooms and dormitories, and from the spaces between friends. And they spring from office cubicles, those most digital and screen-y of places. The element that’s common to all those places is the body that put it into some human context.
AI is now generating faces and voices, creating digital “persons” who will say and do things that the living and breathing versions of that person would never do. That scares the heck out of a lot of us. A robot-generated personality could ever be warm enough to be my friend—much less funny enough.
There are Wednesdays, often in my more drab weeks, when the column idea won’t come to me. When that happens, I walk around my world: wander through the rooms of my house, walk outside, stand in the yard, to see if some memory attaches itself to me. I’ve been doing this for most of my adult life, and the wandering has not failed me. I never miss a deadline.
Is everything I write good? Of course not. I’m only human.
Recently, I wrote a column about how much I like Junior Mints candy, and how all the stores seem to be out of them. Not exactly Shakespeare. Could an AI app write about the the fact the stores are all out of Junior Mints? Of course.
But could a robot write that part about how the Junior Mints feel when they melt in my mouth? How I like the rattle sound when I shake the green and white box? How disappointing it’s been to see the shelf empty?
The robots could never write that, because they don’t taste it, or hear it, or feel it. And I hope they never will.
About the author: Beth Dolinar is a writer, Emmy-award winning producer, and public speaker. She writes a popular column for the Washington “Observer-Reporter.” She is a contributing producer of documentary length programming for WQED-TV on a wide range of topics. Beth has a son and a daughter. She is an avid yoga devotee, cyclist and reader. Beth says she types like lightning but reads slowly — because she likes a really good sentence.