I recently spoke to a college English class about writing, after which I fielded the inevitable questions about ChatGPT. In case you’ve missed all the hubbub, ChatGPT, according to the website gizmodo.com, is an “artificial intelligence tool that allows a user to generate original text. You can ask it questions, give it creative prompts, and use it to generate a whole bunch of different stuff—from poems, to songs, to essays, to short stories” The technology is already being added to Microsoft’s 365 software, under the brand name “Copilot,” and will be embedded into Word, Excel, PowerPoint, and Outlook. In other words, it will soon be available virtually everywhere.
Like most writers, I’ve been giving this some thought ever since the story broke, so I was prepared for the inquiries, which spanned everything from plagiarism to loss of job opportunities. Here’s what I told the young learners:
- As with other breakthroughs, from electricity and nuclear energy to TV and the internet, technology is neither good nor evil. It’s what we choose to do with it that creates the moral dilemmas.
- Based on the results I’ve seen so far, ChatGPT’s output is mediocre at best. Of course, we can’t count on it staying that way, and in reality, AI is bound to improve at warp speed. But at the moment, I’d categorize the writing at middle-school level.
- For probably 80% of the public, mediocre is good enough. For one thing, they don’t know the difference between okay writing, good writing, and great writing. Don’t believe me? Read a page or two from bestsellers like “Fifty Shades of Grey,” “The Da Vinci Code,” or almost any celebrity tell-all. For another, above-average writers have been getting replaced for years by offshore hacks, especially in the area of business communications such as website content, blogs, annual reports, etc. Again, most folks don’t know the difference. And they certainly aren’t willing to shell out extra for it.
- Ultimately, it doesn’t matter. The remaining 20% of writing consumers recognize quality and are willing to pay a premium. Same with any other creative endeavor — painting, sculpture, music, film, graphic design, architecture, you name it. Want to make sure you don’t get aced out? Up your game.
- That goes double for professions such as education, law, medicine, etc., which seem especially concerned about the negative impact of AI. As well they should. For decades, if not longer, the worst examples of muddled, perfunctory, downright painful writing have come out of those disciplines. Maybe there should be less hand-wringing among these practitioners and more respect for the craft of clear, concise communication.
- Innovation being what it is, enterprising programmers will invent software to stop AI in its digital tracks, similar to the never-ending cat and mouse game played by computer hackers and cybersecurity specialists.
- No matter how quickly ChatGPT (and other AI entries) manage to improve, what are the limits? Will it develop the ability to write with wit, style, flair, originality, point of view? What about scintillating dialogue that crackles with electricity? Insight into the human experience? The inscrutable magic that causes us to laugh. Or cry? Will it produce the next Picasso, Michelangelo, Shakespeare, Frank Lloyd Wright, George Carlin or Lennon/McCartney? Or is it destined to hit the wall as nothing more than a capable copycat?
- To write like humans, will bots need to think like humans? Become more human? I’ve seen that movie more than once. I know you have, too. If that happens, we’ll have more to worry about than a loss of revenue.