The Prose and Cons of AI Writing

Robots and sideways writing.

Source: Image courtesy Scott Eberle (2024)

A mischievous friend recently asked whether I would mind if he tasked one of the artificial intelligence (AI) utilities to mimic my writing about play. He had been pestering me to investigate the naughty adult game Cards Against Humanity. We had played it one evening with our partners and two other couples with hilarious results.

He wanted to know whether we could tell the difference between robot writing and the real thing.

So, yes, great question, I was game to play his game. After all, AI has revolutionized endeavors as varied as customer service and weather forecasting. These programs are so smart that they bested Ken Jennings, the Jeopardy champion, in a trivia contest. Even the composition you are reading now has been eased by machine intelligence that helped cut and paste text, correct word agreement, and identify synonyms.

The deep-fake results of the chatbot challenge rang with an uncanny familiarity. I knew of the trouble that editors of scientific journals have encountered with unoriginal computer-generated texts. Still. Maybe the machine could save me some trouble.

But are we at ease with these revolutions?

Can We Trust the Trend?

On the one hand, speculative fiction has delivered a succession of cyborgs and rogue computers, a cohort spearheaded by HAL 2000, the murderous AI. Arthur C. Clarke’s imagination was filled with awful warning. On the other hand, futurists see a great big, beautiful tomorrow. Ray Kurzweil, the computer scientist and inventor, predicts that the “convergent, exponential technological trends are leading to a transition that will be utterly transformative for humanity.” Not the least of these will be the melding of artificial and human neural networks: the holy and magical result, “spiritual machines.”

Writing with Spirit

A thick literary stew nourishes writing. Parsing play, for me, demands that the medium should inform the message. And so, sniffing out my own inspirations, I detect a whiff of other, wittier essayists. A molecule of Voltaire here, a dash of S.J. Perelman there, a pinch of Fran Lebowitz, a smidgeon of Douglas Adams, a sprinkling of Christopher Hitchens and Carl Sagan, more than a touch of Rod Serling’s incisive introductions and afterwords to Twilight Zone episodes, aromas of George Orwell, Jacob Bronowski, Murray Kempton, and Garry Wills. Add to those ingredients a generous measure of Stephen Jay Gould, who once noted that the ritual cannibalism that attended the mating of praying mantises was “a consummation devoutly to be wished.” Try topping that, chatbot.

My most valuable training, aside from the habit of leaving stacks of insistent books open and face down on desks and side tables, was in writing wall label copy for more than a hundred museum exhibits. Though those mini-texts amounted to many thousands over time, by a decree lost in the misty past, none could exceed 65 words.

The reason for this spare discipline, a kind of haiku? Museumgoers’ attention typically divides as memorable objects draw them and interactive installations invite them. Then, too, visitors will carom freely, like pinballs. If the linear written word had a prayer of competing in this beguiling 3D-ecology, the writing needed to be provocative, informative, clear, and winning.

Putting a Chatbot to the Test

The chatbot impersonator delivered a short article, “Cards Against Humanity: A Playful Examination,” by Scott G. Eberle, Ph.D.

The trouble starts with the presumptuous title. Shouldn’t readers decide whether the article will be playful? (Are you likely to believe the flashing highway diner sign that reads “Fine Food!”?) Then, are examinations ever playful? Let’s try adding verve: “A Naughty Night with Cards Against Humanity.”

Then on to the first sentence, first paragraph, according to the chatbot:

“The card game "Cards Against Humanity" (CAH) has captured the interest of millions with its irreverent, often controversial content. It thrives on pushing boundaries and eliciting reactions, be they laughter, shock, or discomfort.”

Artificial Intelligence Essential Reads

AI Chatbots for Mental Health: Opportunities and Limitations

Is "Digital Thought" the Next Cognitive Domain?

This sentence gets an A for effort and a C- for appeal. “Captured the interest.” Sounds clichéd. (Writers should avoid clichés...well, like the plague.) The abbreviation—(CAH)—too inside. What about “controversial”? Are players looking for controversy? The game is mischievous. Wicked. Impish. On to “be they.” Sounds pompous. Thuds like a Puritan sermon. Another word choice swings wide. Do we push boundaries? We cross them, and in this game, we transgress them. So. Let’s try: “When couples, loosened by mojitos, settle in to play Cards Against Humanity, they are in for a wicked evening.” Shorter. Grabbier? Better?

Then the bot writes: At its core, this game isn't just about humor or obscenity; it's a profound manifestation of play, exploring the delicate balance between societal norms and personal boundaries.

Questions. Does the game have a core? Also: Can a game that obliges players to repeatedly say “clitoris” in mixed company aspire to profundity? Unsurprisingly, the robot’s text struggles with the mischievous nature of play itself.

Then the dependent clause “exploring the delicate balance between societal norms and personal boundaries” shows some gravitas; maybe even promise. Except for this: Cards Against Humanity isn’t about balance. Just the opposite, in fact. Transgressions make the game funny and edgy and surprising. Finally, there is nothing “delicate” about the game. It pokes at players, like a poke in the ribs.

The succeeding paragraphs in the machine-written simulation suffer from similar symptoms of writing without spirit and feeling and are deficient in shading, tone, and nuance. Looking at the prose with my name attached, I felt a bit conned.

Bloodless Writing from the Robot

Writing genuinely and simply and accessibly takes care. It takes practice, too. Decades. There is no getting around it. Red Smith, the veteran sportswriter whose prose is so easy to admire, was once asked if he found writing hard. His sentences read so effortlessly. Not hard at all, he is said to have deadpanned, “You just simply sit down at the typewriter, open your veins, and bleed.”