You have /5 articles left.
Sign up for a free account or log in.

A white microwave with two large knobs.

Those of us who write, assign writing and teach writing know ChatGPT will change things. But should we prepare for a major shift in culture and communication akin to the impact of the printing press? Are we mostly going to see changes in how we teach and learn writing, similar to how widespread availability of cheap calculators impacted math education? Should we expect upheavals in how people work and get paid for their work like those brought about by the transformation from agrarian to industrialized economies?

Much of the discussion around ChatGPT in popular media and even among educators predicts extinction. Maybe we’re about to see the end of the essay or of high school English classes or of human-produced communication.

I have been teaching academic writing and helping a range of students develop their writing projects for years. I currently serve on a university committee trying to come up with resources and recommendations about teaching and learning with artificial intelligence tools, including ChatGPT. And, after a lot of reading and thinking, my predictions are based on a much more modest analogy.

ChatGPT looks to me like a microwave oven.

As a Gen Xer, I have a vivid memory of the winter afternoon in the early 1980s when my parents brought home their first microwave. After unboxing this futuristic cooking device and hefting it onto its special cart, my dad read the manual. Our microwave oven was large and heavy, with wood-grained paneling on the top and sides. Knobs allowed one to adjust cooking times and power levels. My younger siblings and I were fascinated by the fact that when you turned it on, the small window in the door lit up.

We spent a few months delighted by our newfound capacity to cook bacon without a frying pan and to heat canned chili without turning on the stove. We learned a microwaved marshmallow would expand to impressive proportions and a hot dog might explode. We felt surprised and maybe a bit disdainful that our grandmother—an excellent home baker—didn’t see the need to buy this wonderful new appliance. Other folks my age likely can recall similar experiences.

This flight of nostalgia has me seeing what I think are key similarities to the latest technology so many of us are playing around with, ChatGPT.

  • When home microwaves became widely available in the 1970s, they were presented as new, modern, even futuristic. But microwave cooking wasn’t entirely new. The technology had been developed in 1945. Earlier microwave ovens had been used on planes and in restaurants for decades. ChatGPT is similarly an extension of tools lots of people have been using without much concern: grammar correction programs and autocomplete prompts and help-desk chat bots also rely on algorithms that analyze large data sets drawn from real-world language use. We’ve been living with artificial intelligence for quite some time. What’s actually new is widespread availability—think of it as the home kitchen version of artificial intelligence.
  • My family’s first microwave oven got presented as potentially dangerous. Parental warnings not to stand within six feet of the door and grandparents’ worries that microwaved food might be irradiated now seem silly. (We did learn quickly, however, that microwaving a whole egg or an unpierced potato would cause a mess.) The more mundane dangers are now well understood. Because metal in the microwave creates sparks that amount to a fire hazard, dishes and kitchen tools now come with labels indicating whether they are safe for microwave use. Superheated liquids can splatter unexpectedly and cause burns, so recipes and some food products now include instructions for stirring. We’re now at the stage where examples of AI chat bots doing outrageous things are being trotted out—telling a journalist to leave his wife or sharing a recipe for napalm. Those seem like real risks. The end of human-produced writing, however, seems as unlikely to result from ChatGPT’s existence as a cancer pandemic caused by microwaved food. While reasonable people can worry that a new technology is dangerous, it will take time to figure out what risks are real and then to work out some safeguards.
  • Most notably, the excitement about home microwave ovens led some to believe everything about home cooking would change. Look at any used bookstore’s cookbook section for a sense of what “experts” were promising in their microwave recipes in the 1970s and ’80s. Sure, you could use your microwave to boil water or reheat leftovers. But if you were really skilled, you could use it to produce a whole Thanksgiving dinner, including a beautifully browned turkey and crisp-crusted apple pie. Titles like Microwaves Are for Cooking and The New Magic of Microwave Cookbook recommended using your new appliance to steam a whole lobster and to make peanut brittle. Complex recipes suggested fiddling with power levels, opening and closing the door frequently to add ingredients, and even stocking up on unusual products that could simulate browning or smokiness or crunch. Similarly, there are now rumblings that ChatGPT will produce high-quality writing ranging from student essays to long-form journalism, from weather reports to movie scripts. Suggestions for how to interact with the program to refine instructions are presented as a surefire way to generate better arguments and turns of phrase. So far, though, the ChatGPT text I’ve seen generated in response to even the most refined prompts is the equivalent of a microwaved chicken breast. It’s cooked, but not what anyone would point to as ideal. It might be possible to get good outputs from ChatGPT, but we will likely discover that achieving a lot of those results will be more work than writing the old-fashioned way.
  • Since those heady days when we got our first home microwave ovens, most of us have figured out they are suitable for some tasks and bad for others. In addition to using a microwave to warm up soup, I find it the perfect tool for melting chocolate and butter. Award-winning chef David Chang reportedly perfected a microwave recipe for chawanmushi, a Japanese custard dish. Stephen King advocates for a microwave technique that produces perfectly cooked salmon fillets. My prediction is that ChatGPT will eventually be used in similarly varied ways. Some highly skilled people (including those who already know a lot about writing and research and communication) will find a handful of unexpected but effective applications. Most people will discover one or two things it’s good for and consider those to be reasonable shortcuts.
  • In the decades that microwave ovens went from a newfangled kitchen appliance to a ubiquitous tool, there are ways that the microwave changed how people eat at home. The freezer aisle of any grocery store is full of entrees designed to be cooked or heated in a microwave, and there are people who occasionally or habitually rely on the convenience of these items. A couple of generations of people now find it unremarkable that even a child can use a microwave to prepare instant oatmeal or a bag of popcorn. But people still cook. One might even argue that the widespread availability of low-effort, often mediocre food (including microwavable entrees) led to more interest in project cooking at home and to greater appreciation for the talents of fine chefs in restaurants. ChatGPT might well have a similar effect. If much of the boilerplate text of form letters and weather reporting gets generated by ChatGPT, the differences between these dull but useful forms of communication and, say, the surprises found in a human-generated love letter or sonnet about a summer day could become easier to see—and more widely appreciated.

Even as I muse about how current discussions of ChatGPT parallel popular advice about the promise, use and dangers of home microwave ovens in the late 1970s and early 1980s, I admit this might not be the right analogy. Maybe ChatGPT is more like the steam engine. Or polyester. Or antibiotics. In any case, we should keep striving for analogies. To date, that’s one of the things human beings are great at that algorithms still struggle to do well.

Some of the comparisons we trot out will be half-baked, while others will be overcooked. Through thoughtful experimentation and skepticism about both the most optimistic and most doom-filled predictions about ChatGPT’s likely impacts, we are likely to wind up finding this is one more tool we can live with.

Erin E. Kelly is an associate professor of English and director of the academic and technical writing program at the University of Victoria in British Columbia, Canada.

Next Story

Written By

More from Views