Sign up for The Key Point of View, our weekly newsletter of blogs and podcasts!
Unless you’ve been living under a rock for the past few months, you cannot fail to notice all the headlines around a certain next-level artificial intelligence (AI) engine and its impact on content generation. ChatGPT—a voluble, relentless, and potent chatbot—has made a lot of people sit up and take notice.
Released as a free-to-use tool in November 2022 by San Francisco-based tech company OpenAI, its use is making speedy inroads into the world of academia, literature, and even scientific publishing. In fact, Nature recently reported that ChatGPT has landed at least four authorship credits on published papers and preprints! An earlier article spelled out the difficulties around its use, with some cases where researchers were unable to differentiate between abstracts written by the chatbot and those by actual flesh-and-blood scientists. It’s a compelling example of the influence AI-generated content will have in the future, showing just how powerful it (and those inevitable smarter successors to come) can be.
Predictions about the death of writing by the human hand may be wild, but there’s something quite unpredictable about where the use of ChatGPT is heading right now. Some say it’ll give Google a run for its money because it’s rather good at providing answers to queries. One of my friends (incidentally a journalist and best-selling author) obtained the definition of quantum physics in Sumerian in a jiffy.
ChatGPT is very good at next pattern matching—that is, coming up with its next sentence based on words that have come before—and is very capable of coming up with countless original combinations of words, so it’s definitely suited to academic work. It will be more difficult for teachers to spot and prove plagiarism. In fact, lecturers at UK universities are being urged to review the methods of assessment in certain courses, with a possible return to more traditional pen-and-paper based essay writing in some subjects. Schools in New York, on the other hand, have banned the use of ChatGPT entirely.
Recently, OpenAI has responded to educators’ concerns by releasing a new “classifer” tool that has been trained to detect whether text has been written by an AI bot (any bot, not just ChatGPT), but warns it’s not completely reliable yet. Open AI researchers said that, while it was “impossible to reliably detect all AI-written text,” good classifiers could pick up signs that text was written by AI. The tool could be useful in cases where AI was used for “academic dishonesty” and when AI chatbots were positioned as humans, they said.
As a bard and lyricist, ChatGPT may have some way to go to match the high art of Tennyson and Dylan. Nick Cave is not impressed. A fan sent song lyrics to him that he had instructed ChatGPT to write in the style of Nick Cave. The resulting effort was filled with brooding biblical imagery but is, in the words of the elegant man himself, “a grotesque mockery of what it is to be human.” Still, there are plenty of admirers.
My friend the author’s long view leans towards the pessimistic side of the fence. To paraphrase: “…The verdict is bad, sad, and terminal. Five thousand years of the written human word—and 500 years of people making a life, a career, and even fame out of those same human words—are quite abruptly coming to an end.”
I’ve been eager to try ChatGPT out for myself, but I’ve been thwarted thus far with the site showing an “at capacity” notice. Following the incredible explosion in interest, traffic to the site is heavy, and it has been known to crash. Add on top the high running costs and its little wonder why OpenAI appears to have put limits on usage. I can join a queue, leaving me with plenty of time to think about what I want ChatGPT to write. Perhaps the end to this piece. My prompt can be “Will ChatGPT cause the destruction of writing?”
I think not. I believe ChatGPT cannot replace the human hand (never mind what Buzzfeed thinks) as it isn’t an expert, isn’t an editor, doesn’t do nuance or context, or can achieve the layers of complexity and analysis that a journalist or a skilled author can craft. And of course, ChatGPT and the bots to come rely on digesting, rearranging, and regurgitating existing content. If we get to the point where humans stop generating that original content, all the AI bots will wind up just retreading existing material in an endless echo-chamber loop…which is not unlike the Tom Clancy canon.
But where it will be particularly potent in the future is when businesses train it on their own knowledgebases and the program becomes more specific and personalized to particular industries. And like other AI technologies, such as robotic process automation (RPA), it can give humans a head start by taking over the tedious “gruntwork” aspects of a task to free up knowledge workers for more critical tasks…those that require a human.
Log in to the InfoCenter for more information about smart workplace trends in our Office CompleteView Advisory Service. If you’re not a subscriber, just send us an email at firstname.lastname@example.org for more info.