Back when I was seven years old, I wrote a story about Glinda the Fish. I don't remember anything about it beyond a few vague watery adventures Glinda had. Certainly the story rambled, had no story arc, and lacked a conclusion, though I did include crayon illustrations of a very pretty and feminine fish decked out in glittery scales. Hey, I was seven.
Obviously the writing bug bit early, and I've enjoyed practicing the written word ever since.
Perhaps for this reason, the artificial-ness of writing done by artificial intelligence drives me nuts. I'm complained about this several times (here and here).
Oddly I'm conscious of times I use turns of phrase that might be misconstrued as written by AI. I submitted a magazine article recently, for example, concerning the honey hunters of Bangladesh (people who harvest wild honey in the coastal areas of that nation). In the second paragraph, I set the scene by describing what the mangrove forests are like, which included this line: "It is a moist, humid, and unforgiving ecosystem that straddles land and sea, and protects the coast from erosion during high tides, tsunamis, and cyclones."
It was the term "straddles land and sea" that I almost took out because it sounded like AI, even though it describes perfectly the placement of mangrove forests.
Similarly, one line from our recent trip to Italy also triggered my AI radar. In this blog post, I described some Roman ruins we saw, including the impressive brick work, as follows: "The fact that the brick work has held up so splendidly for thousands of years is a testimony to the skills of Roman craftsmen." It was that word "testimony" that tipped the line over into the realm of sounding artificial. It's bugged me ever since.
Every writer develops his own unique voice when mastering the craft, and I guess you could say the same applies to artificial intelligence. In a piece entitled "The Rise Of AI Writing And The Decline Of Human Voice," it states:
"New research suggests that widespread use of large language models is
making language more uniform. A study conducted by University of
Southern California found that after the release of ChatGPT, diversity
in writing styles declined across several forms of communication,
including scientific publications, local journalism, and social media
posts. Researchers observed fewer differences in vocabulary choices and
sentence patterns, pointing to a growing preference for polished,
formulaic language. ...
"Researchers at the Max Planck Institute for Human Development analyzed more than 740,000 hours of spoken and written material and found that certain words commonly associated with ChatGPT responses are appearing more frequently in everyday communication. Words like 'delve,' 'meticulous,' 'boast,' and 'comprehend' have become increasingly common, suggesting AI-generated language may be shaping human speech habits as well. ...
"Not everyone sees that as progress. Alex Mahadevan of the Poynter Institute for Media Studies argues that AI-generated content often feels empty despite being technically sound. He described it as noticeably 'soulless' and 'mediocre,' adding, 'There's no art in it.'"
Or, as I'm fond of stating, AI writing is a lot of "blah blah nothing."
At this moment in time, we are straddling (that word!) the cusp of human-generated writing and AI writing. But the handwriting is on the wall, and the latter will become the norm. Not only is it easier (give a command, and hey presto! your essay or advertising copy is written), but the younger generation will grow up unable to discern between the real and the fake ... especially since AI shows every indication of becoming indistinguishable from the human voice. Personally I find that very sad.
I don't know how long I'll be able to cling to a career as a writer, but I'll enjoy it while I can.
No comments:
Post a Comment