How Playing Chess Conquered My Fear of ChatGPT

By Ryan Weber

Whenever I panic about how ChatGPT will affect my opportunities as a writer, I calm down by remembering that I’m a mediocre chess player.

I’ve loved chess since my dad taught me to play in elementary school. I’ve played 1610 games on chess.com. I watch YouTube videos and read books on strategy and openings. I purchased chess lessons from the silent auction at my Unitarian church. After all of this effort and experience, I have become… a perfectly competent chess player. Currently, 249,915 people on chess.com have higher rankings than my 1289 rating, and I could never beat many of these prodigies, wannabe grandmasters, actual grandmasters, club champions, and Kenyan high schoolers. I certainly can’t beat the best chess AI. Neither can the greatest human competitors. Deep Blue famously defeated then chess grandmaster Gary Kasparov in 1997, and AI has since evolved to the point where chess engines like Stockfish and AlphaZero always triumph over top-rated humans. 

Yet people still play chess for its profound mental engagement: puzzling over the board, imagining various moves to find the greatest chance at checkmate, rejecting or sometimes embracing bold gambits, intuiting your opponent’s strategy. Chess promises that by engaging in what Vladimir Nabokov calls a “long, long interval of thought,” a player can deduce “a bewitching, brittle, crystalline combination” of moves that might spell victory or disaster, depending on the opponent’s response. I derive immense satisfaction in finding the perfect fork that breaks an impasse or devising an elegant escape from a seemingly hopeless position. 

Chess is intense, frustrating, exhilarating, and exasperating. None of this engagement gets diminished because AI can deduce in nanoseconds the solution I agonize, and often fail, to find. I don’t want the easy answer. The human struggle is part of the satisfaction. Ultimately, AI’s perfect chess is boring. 

This returns us to ChatGPT, the impressive natural language processing AI tool that can write pretty impressive passages. Admittedly, ChatGPT sometimes generates lifeless, formulaic, inaccurate, or even comically bad output. When it comes to writing ability, ChatGPT is not Stockfish or Deep Blue, or even Deep Thought, the more rudimentary computer that beat grandmaster Bent Larsen in 1988. ChatGPT can’t outwrite Colson Whitehead, Zadie Smith, or Jhumpa Lahiri. If writers were ranked like chess players, ChatGPT would earn around a 1000 rating. Perfectly competent in some situations, but not especially adaptive, creative, or compelling. 

Still, AI writes better than many humans. Its rapid improvement implies a future where AI may write more accurate, beautiful, compelling, moving, and engaging prose than any human being. In two, five, ten years, AI will likely compose or assist with most of the writing that institutions need, like company newsletters, social media content, legal briefs, and many other boilerplate documents. This could mean catastrophic consequences, or at least massive changes, for many professional writers. 

Regardless of how far algorithms advance, we will never cede the entire realm of writing to AI, just as people have not stopped playing chess despite the superiority of computers. We will likely never award a Nobel, Pulitzer, or screenwriting Oscar to an AI writer, in the same way we’re not going to crown AlphaZero grandmaster. We want to watch humans play chess. And in many circumstances, we want to read what humans wrote, knowing that humans wrote those works. Not in spite of the imperfections, but because of them. We want human authors’ insights, their experimentation, their voice, their hard-won experience. 

Writing is not a perfect analogy to chess. It’s more easily monetized and more societally ubiquitous, and chess offers clearer victories and definitive logical precision. Regardless, writing offers similar intellectual engagements to chess. Trying out multiple variations of a joke and finally crafting the perfect turn of phrase. Concluding an essay and uncovering insights that weren’t part of the original concept. Falling into a flow state where time fades away in moments of glorious concentration. Writing is intense, frustrating, exhilarating, exasperating. As Hemingway observed, “Writing is something that you can never do as well as it can be done. It is a perpetual challenge and it is more difficult than anything else that I have ever done—so I do it. And it makes me happy when I do it well.”

Of course, like chess, writing gets frustrating and often results in failure, but we do it anyway. Jamaica Kincaid describes the sadness of the blank page, of seeing a sentence “just sitting there followed by nothing, nothing and nothing again.” But I don’t want to turn the blank page over to a machine. I want the whole process, not just the finished product, because the frustration makes the effort rewarding. 

However, I’m being unfair by positioning AI chess engines and ChatGPT as antithetical to human creativity. As a mediocre chess player, I use AI daily to improve my game. I run the chess.com algorithm on completed matches to identify my few great moves and my many, many blunders. In these cases, AI allows me a second chance to replay the move and to hopefully learn from my mistakes. 

Writers make, and learn, from blunders as well. As Toni Morrison remarked, “I don’t mind writing badly for a couple of days because I know I can fix it–and fix it again and again and again, and it will be better.” AI writing tools may play a similar role to chess engines, helping writers spot their weaknesses, identify better options, break through a block, up their game. In the best case, AI will serve as a brainstorming and revision tool for writers, much like templates, sample documents, and editorial review do today. 

After his loss to Deep Blue, Kasporov moved past his bitterness and introduced the concept of Advanced Chess, which unites humans and AI for greater performance. Human/computer partners play each other, uniting the precision and predictive power of AI with human intuition and theory of mind. As Kasporov told the MIT Technology Review in 2022, the future of chess lies in “finding ways to combine human and machine intelligences to reach new heights, and to do things neither could do alone.” 

The future of writing will likely follow a similar path. Humans will still often write alone, AI will dominate humans in some writing endeavors, and people and AI will sometimes work together in productive writing collaborations. Ideally, these collaborations will involve transparency about when and how AI is involved. Readers will always want to know if AI assisted with writing that has a byline, just as humans want honesty when their human opponent gets help from a computer (or bluetooth-enabled anal beads, a story worth Googling if you don’t know it). 

ChatGPT feels like a genuine gamechanger, a Pandora’s box that we cannot close. It will change writing forever, perhaps even more drastically than AI changed chess. Not all of the changes will be good news for writers. But chess and writing engage distinctly human parts of us, and we will never get quite the same pleasures from ceding these realms to AI, even as we bring AI onboard as a partner in the endeavor. 

Ryan Weber teaches writing at The University of Alabama in Huntsville. His writing has appeared in McSweeney's, The Conversation, Slackjaw, and several nerdy academic journals.