Member-only story
Critique of the AI Revolution
Recently I decided to see what all the Hoopla is about with ChatGPT
I downloaded the AIPRN plugin for Chrome and gave it a task. The results were pedestrian at best. It paralleled the slow decline in written language and comprehension by the general public.
I prompted the AI to write an article about my photography collection and the platform it is on. The system seemed to know what ClickaSnap.com is. It was described quite accurately and with glowing praise.
My user account, Lens4anEye, was completely misinterpreted. The article was written in the First Person Plural as in “At Lens4anEye we strive… “etc etc. The references were such that I seemed to be an entire team who would identify, select and promote exceptionalism of highly qualified photographers around the globe.
As a random person being presented with the content thusly created I would have no idea that most of the text was “boilerplate” language there but to fill column space.
I was reminded of my 5th Grade writing assignments where I was supposed to use at least 3 references to write a factual paper on some subject. I always liked to write about Natural Resources. Hence my collection of The Natural Resources of.. Brazil, Argentina, Columbia, Spain et al. My “training sources” were the Encyclopedia Brittanica, Comptons and maybe National Geographic. My biggest concerns were to meet the length requirement and of changing the wording sufficiently so as to not appear to have been merely copied. I was never never concerned with sounding repetitive or wordy.
That half century ago was a time for learning to write at all. There was a required level of reading comprehension necessary to meet to assignment with both content and grammar. I suspect I could obtain an equally pedestrian report using ChatGPT with the prompt “summarize the natural resources of Brazil in a 5 page report.” The “sounding like a 5th Grade boy” part would be optional.
The great success of AI text generation is predicated on the reader having a low comprehension level, or at least one comparable to the AI writing skill.
I’m thinking AI would completely fail at writing comedy, puns, double entendres, satire and parody. For that…
