I was at a Christmas celebration last week with a bunch of Rotarians (my wife is one of the spinners. I haven’t been one since 2013) when the subject of what we do for a living came up.
Aside from finding out the husband of one of my table mates is a fellow connoisseur of progressive rock, I also found that three of the four people at my table (including my wife) are superfans of ChatGPT.
Having been a news writer since the mid-80s, any discussion that veers into the AI lane gives me, to quote superhero Buffy Summers, “the squick,” particularly when it comes to writing.
While I can see that not all people are wordsmiths and need help from time to time, what worries me is that the people supposed to be professional are more and more diving into the AI pool, and mostly into the deep end.
GIGO, by the way, was a phrase I learned moons ago in reference to writing computer programs (while dusty, my geek cred extends to writing Basic and Fortran code, and in later years to HTML). It means garbage in, garbage out.
For instance, it came out recently that in a story by the Washington Post regarding Secretary of War Pete Hegseth, which was sourced anonymously (go figger), that an order to destroy all survivors of a boat carrying narcotics was given by SecWar. The salacious story had these “sources” quoted as saying that the man himself, after the boat was sunk, ordered further missile strikes because there were two survivors floating in the water.
A later story, by equally editorially corrupt competitors the The New York Times, found that the Post story was garbage. Garbage in from the sources, garbage out from the reporting staff.
Later in the week, I came across a story that found that a claim by the climate police that rising CO2 levels will mean the world will be $38 trillion poorer was full of, well, hot air.
The study, which was released last year, was widely reported by mainstream media outlets without the benefit of any verification. Every big-time news hustler doomcried the results of climate change as another reason why we all had to do our part to stop this insidious event.
Turns out that the research was extremely flawed; stories said the basis for the results cited in the study was based on “implausible” (read: garbage) emissions scenario.
Garbage research in, garbage reporting out.
There are I’m sure numerous examples of the same thing across all levels of news. I recently read a piece from a colleague and former boss who said that he encourages his reporters to use ChatGPT to assist in writing their stories as it makes his editing job so much easier.
So, the premise here is that when your source material is junk, your results are junk.
I made that point to my table mates, with the admonition that these AI bots only dig up what is already out there. What I mean by that is it’s long been admitted that the majority of information found on the internet is suspect, and extremely unreliable. Further, agencies that are supposed to be trustworthy are using unreliable information for their web entries, which muddies the information pool even more.
While that is problem enough, I further posited that unless people are taught the proper way to do things like write or add or even research, not only will any results be suspect, but these same people will not develop any analytical skills that would show them the errors of relying on AI.
My table partners were quick to say it’s only used for writing the occasional email or transcribing notes or some other innocuous task, but as with everything else, once that first step is taken, it’s not too long before the sprint.
Tony Farkas is editor of the Trinity County News-Standard and the San Jacinto News-Times. He can be reached at tony@polkcountypublishing.com.
