Introduction
I was questioned about the AI-generated writing in my last publication, "3 Years of Laravel Jobs: What 699 LaraJobs Emails Actually Say About the Market," and it made me think.
Even though I am not a native English speaker, I created this package and documented it 6 years ago; I believe that AI could have helped me improve my grammar and reduce typos. It seems pretentious to think that AI couldn't have done a better job documenting how to use it, as I understood what I was doing, for whom I was writing, and why.
This is not an excuse for the fact that I used AI in my last posts. I have used the help of AI to write them, even though the ideas and the research were mine. It is tempting to delegate such a task, in which I am not proficient by any means, to a black box that has seen all the blog posts on the internet.
Finding Forrester
I am writing this post not only because of the critics but also because a YouTube reel caught my attention, which led me to rent the movie "Finding Forrester," as it relates to the core of the discussion. In "Finding Forrester," Jamal Wallace, a brilliant sixteen-year-old writer, is questioned about the authorship of his work by Professor Crawford, who once wrote an unpublished book and chose to pursue a career in lecturing, ultimately becoming a barrier that no student could surpass. His frustration turned him into a very rigid tutor whose judgment was so harsh on his students that no one could argue in his classes.
Jamal met William Forrester, which accelerated his already excellent writing skills, and Professor Crawford couldn't take it. A sixteen-year-old Black boy from the Bronx was writing pieces of that quality. "Of course it wasn't you, Jamal," said Crawford.
The Biased Metrics
Crawford's belief that Jamal could never have written those pieces was based on specific words and rules used by AI detectors today, which, while often indicating high accuracy, do so not because they are effective but because the prevalence of AI usage is very high. The number of false positives is high, particularly for texts written by non-native speakers, as revealed by a Stanford article titled "AI-Detectors Biased Against Non-Native English Writers," which links to the paper discussing how GPT detectors are biased against non-native English writers.
Conclusion
Both the critics and the movie made me realize that the initial posts of this blog were genuine, and I like them for that; they were 100% my word, like this text, with no AI intervention. To be fair, the vocabulary I applied in this text was more than I expected, and I know I can use more words. What amazes me is the fact that this is not a technical post.
As Jamal had Forrester's tutoring, I must use AI as such instead of a ghostwriter, but remember that reading and practicing the art of writing is what it takes to achieve mastery. Jamal didn't need Crawford's validation. Neither do I.