Have you noticed any change in my writing style? How about my vocabulary or phrasing? Over the past few months I have allowed the web host AI Assistant to suggest improvements to my writing.
Here’s an example. One of the most common AI recommendations is when I generate long, maybe run-on sentences:
Although I am interested in seeing these suggestions, I am sometimes trying to get a specific point across which has its meaning changed by changing the long, rambling sentence structure into shorter, more concise groups of words.
Here’s what it looks like after I allow the AI Assistant to make the changes:
“I am interested in seeing these suggestions. However, I am sometimes trying to get a specific point across. The meaning changes when I change the long, rambling sentence structure into shorter, more concise groups of words.“
Now that wasn’t bad! I like it. I think I’ll use it (I say to myself). However, after making the changes based on its own review, it now suggests the the word “However”, is complex. Thus it suggests this change.
“…these suggestions. But, I am sometimes…”
And I think, “What’s up with that?!” First, why suggest a change that transgresses its own analysis? Second, I happen to like the word “however” better than “but.”
Thus, it is an ongoing struggle to decide whether I should “listen” to the AI or just turn it off and trust my own intuition. Oops! There’s another long sentence! What is the AI telling me now?!
“It is an ongoing struggle. I must decide whether to “listen” to the AI. Alternatively, I could just turn it off and trust my own intuition.“
And now it doesn’t like its own suggestion to use the word “could.” It would rather I just leave out the word “could.” “Alternatively, I just turn it off and trust my own intuition.“
Are we having fun yet?
I’ve read and listened to articles about the efficacy of using current iterations of Artificial Intelligence. Some news organizations are replacing people with AI to generate stories such as (the AI is telling me to use “like” instead of the more complex “such as”) after a high school football game, giving scores, and analysis of the teams’ performances.
(“Some news organizations are replacing people with AI to generate stories after a high school football game. The AI provides scores and analysis of the teams’ performances.“)
In this case the AI changed my intent. I wanted to be more inclusive than just an article about a football game. Leaving out “such as” or even “like” totally changes the meaning.
I remember being on an accreditation committee for the college at which I taught technology courses. My team was responsible for reviewing the state of the college’s ethics and civic responsibility. I was chosen by my team to write our results, analysis, and recommendations to the larger committee for inclusion with the other required areas of self-review. (I’m ignoring another AI recommendation for long sentence.)
The college board hired a retired English professor to review all of the team reports. The goal was to bring them into a cohesive, uniform format. My team reviewed his rewrite of our draft. We clearly saw that he had changed the overall intent of our findings! Protesting, we were called to a meeting with the VP for Academic Affairs and the editor to discuss our disagreement.
I’m no English major. I didn’t even do well in that or any other subject in high school or college. Here I was arguing with these doctoral degrees that they can’t change words and sentence structure just for conformity’s sake. Not if the meaning is lost. Especially when it glossed over some less-than-complementary findings.
I later realized, after winning the day, what I had done. Then I worried about keeping my job! I had butted heads with the VP and expert. Would it end my career? Did I really know what I was talking about? “Holy s**t!”
Nothing bad happened. And, as you can tell, I continue to challenge others’ recommendations. Though designed by experts, I certainly won’t just accept the “word” of a computer over what I want to write!
We’ve all heard the potential value of AI. We’ve also heard of its dangers. When it comes to AI assistance, there is certainly room for improvement. I don’t think we will be replaced completely any time soon. I do, though, continue to remember humans being used as batteries in The Matrix! (A very innovative movie for its time, I might add.)