Sound and fury, signifying nothing
Recently I have found my social media feeds inundated with a panoply of essays enhanced by AI writing assistance.
Those with sentences written in short, staccato bursts. No frills, just simple writing.
These essays are published by all sorts of accounts, from the mass of socialites and entrepreneurs trying to break into the mainstream all the way to the self-styled thought leaders maintaining their public erudition.
ChatGPT premiered in late 2022 with the writing ability of a student attempting to finish their college application essay the night before its deadline. Four years later, the average quality of writing published by these AI-assisted accounts have almost achieved this baseline. It should be highly alarming to readers that even the most reputable accounts, once they start using this AI assistance, revert to the same writing recipes spit out by their AI of choice.
The average sentence has such basic subject-verb-object construction that we would be worried if a child could not write anything more sophisticated. The average paragraph length barely breaches the three-to-five sentence guidance from grade school as if the AI—or the writer themselves—were afraid of the teacher penalizing them for going on at length on their ideas. To say nothing about the various platitudes or aphorisms skimmed from the surfaces of issues that these commentators wish to opine upon, then served to thousands of followers looking to their guidance to navigate the tumult of the world.
What does it say about these authors that they are willing to publish the most basic essay outputs given by their AI, without the editorial intent to reflect the meanderings and wanderings their minds take to develop their insights? What does it say about us, that we have accepted this kind of short-form, summarized content as substitute products for professions of belief and analysis in the marketplace of ideas?
Perhaps this is simply the next step of algorithmically-determined engagement that has been ongoing for at least the past decade. The mid-2010s witnessed the explosion of divisive political content spammed to drive various societal processes, elections foremost, squarely into their favor of equally polarizing personalities. We now see that the content itself was not the point, less so their veracity. They were fixed points for organic and manufactured support to flood across the internet with such volume and frequency as to drown out dissenting opinions. This was a dominant strategy as platforms favored likes, comments, shares, and subscriptions over public goods of civil discussions and a shared grasp of reality.
The turn to short-form content and videos accelerated this trend. Why watch a three-minute video when a 30-second clip exists? Why engage with an essay’s opinion when a paragraph of algorithm-curated zingers is shorter to read and quicker to share? The seconds of eye-catching content compound into the hours of “doom-scrolling” precisely because our attention has been obliterated so thoroughly by this non-committal content as to erode our own accounting of time.
Now we have advances in AI at our disposal. The promise of AI included freeing people up to think deeply about truly intractable problems, confident that AI would have the computational power to design the analytical architecture for sifting through immeasurably large amounts of data to our liking. Instead, it seems we have decided that it would be better to let the thinking machine counterfeit the human mind.
We have Claude and Gemini formulate opinions and beliefs that Juan and Juana should be striving to articulate themselves, so that they are left with the simple task of entering the prompts and then subsequently sharing this content on their feeds. Consider that these pieces eventually drain into the datasets that AI uses for its cognition. We are increasingly likely to see AI-assisted responses to AI-assisted essays, photocopies of photocopies of photocopies until the ink of knowledge is barely rendered.
Why my insistence on people’s responsibility over their authorship? Simply sharing information is an incomplete appreciation of language, else we should surrender all our communication to the computer’s binary signals. It is sharing information that is important to us, as humans attempting to live our lives with each other. It is the challenge to describe our vibrant intellectual, emotional, and spiritual experiences in the right words to each other, and in the right words to ourselves.
This excessive reliance on AI assistance for condensed, viral thought pieces misses the forest for the trees. We now treat what could be shared tragedies and triumphs with fellow humans as data points to improve the AI prompt and social media output. What opportunities we have to express our own languages of principle, belief, and importance, instead are turned into machine-calculated sound and fury, signifying nothing.
—————-
Miguel Ventura, 29, lives in Quezon City.
