Will ChatGPT Take Over My Job as a Writer?
I experimented with the program to see what it could do.
After my daughter was born 14 years ago, I quit my full-time job as a magazine copy editor and started freelancing, picking up projects from various publications and writing or editing everything from recipes to blog posts to full-length features. My bread and butter, though, was writing how-to articles for publications like Mommy Nearest and This Old House, where I would be assigned a topic to research and then write about with authority.
I made a decent wage producing this type of practical content, and it sustained me during a time when my kids were young and I needed my job to be flexible. I currently have a full-time job again, but for more than a decade, I was grateful to be able to rely on that specific type of work for extra income.
Today, the advent of ChatGPT has completely obviated that category of writing. How-to writers are our generation’s milkman — once performing a service that was very much in demand, and then suddenly rendered completely obsolete by societal advances. This became clear to me last week, when I finally got a chance to try out the program. I signed up for a free account (in exchange for my email and phone number) and gave this AI thing a whirl.
The interface consists of a single chat bar, into which you can type anything you want: a question, instructions, a statement. The AI will reply accordingly, mining the content of the Internet to curate a response that simulates human conversation.
I started off trying to compare an article I’d written for This Old House called “Home Library Ideas for Every Space” with a ChatGPT prompt on the same subject matter. Less than five seconds after firing off the topic, the program responded with the following:
Here are home library ideas that can work for different spaces:
Cozy reading nook: Transform a small corner into a comfortable reading nook with a bookshelf or floating shelves above a comfy chair or bench. Add a lamp and a few decorative items to create a cozy atmosphere.
Built-in bookshelves: For a more permanent and polished look, consider built-in bookshelves that can be customized to fit the space and style of your home.
Color-coded bookshelves: Organize your books by color for a unique and visually appealing display.
Floor-to-ceiling bookshelves: Make a statement with a grand, floor-to-ceiling bookshelf that showcases your collection of books and decorative items.
Library ladder: Incorporate a sliding library ladder for easy access to higher shelves in a tall bookshelf or for a touch of vintage charm.
Hidden bookshelf door: Create a secret room or hideaway with a bookshelf door that blends seamlessly into the wall.
Vertical bookshelf: Maximize space with a vertical bookshelf that can be mounted on the wall and rotated for easy access to books.
Bookshelf as room divider: Use a bookshelf as a functional and stylish room divider that separates living spaces while still allowing light and air to flow through.
Window seat bookshelf: Combine seating and storage with a window seat bookshelf that provides a cozy spot to read and store your books.
Wall-mounted bookshelves: Make the most of a small space with wall-mounted bookshelves that don't take up any floor space.
Floating bookshelves: Add a touch of whimsy with floating bookshelves that give the illusion of books magically floating on the wall.
It’s eerie how closely this answer, produced almost instantaneously, matches the article I spent a few days painstakingly researching, writing, and editing. It’s not a replica, but it’s definitely close. Worst of all, I realized that the tone of the AI-generated answer was the same conversational, useful, and encouraging tone I attempted to convey in my own article. (Was it doing a great job of simulating a human voice, or had I been doing a great job of simulating a computer-generated auto-response?) ChatGPT’s version of “Home Library Ideas for Every Space” wasn’t as long or as detailed as mine, but with a bit of fleshing out, it could be virtually indistinguishable.
With this confirmation that the jobs of practical writers will soon be completely taken over by artificial intelligence, I decided to explore fiction writing. What would ChatGPT say if I asked it to “write a mystery novel starring an Asian-American protagonist ” — which is the broad strokes of a story I’ve been mulling over recently? Here’s the response, again produced in under five seconds:
Detective Jane Lee was one of the best detectives in the San Francisco Police Department. As an Asian-American, she faced her share of discrimination and prejudice, but her sharp mind and determination had earned her the respect of her colleagues and the trust of the community she served.
One day, she received a call about a missing person case. A young woman named Mei Ling had disappeared from her home in Chinatown, and her family was desperate to find her. Mei Ling was a promising student at a local college, and her disappearance had left her family and friends bewildered and scared.
Jane immediately sprang into action, interviewing Mei Ling's family and friends and combing through her social media and cell phone records. She found no signs of foul play, but something about the case didn't feel right. It was as if Mei Ling had vanished into thin air.
As Jane dug deeper, she uncovered a web of secrets and lies. Mei Ling's boyfriend had been cheating on her, and her parents had been hiding their financial struggles from her. But none of these revelations seemed to explain her sudden disappearance.
One night, as Jane was reviewing the case files in her office, she received a mysterious phone call from someone claiming to know what had happened to Mei Ling. The caller instructed her to meet at a secluded park in Chinatown, and warned her not to involve anyone else.
Jane knew it was risky, but her gut told her that this could be the break she needed. She arrived at the park and waited, her heart racing. Suddenly, a figure emerged from the shadows, holding a piece of paper. It was a ransom note, demanding a large sum of money in exchange for Mei Ling's safe return.
Jane knew she had to act fast. She alerted her team and set up a sting operation, tracking down the kidnappers and rescuing Mei Ling before any harm could come to her. It turned out that Mei Ling had been targeted because of her family's business connections, and the kidnappers had hoped to extort money from them.
Thanks to Jane's quick thinking and determination, Mei Ling was safely reunited with her family. And as for the kidnappers, they were brought to justice, thanks to the efforts of one intrepid detective who refused to give up until the case was solved.
As I read over ChatGPT’s hastily yet confidently written story, I felt a mix of awe and relief. On the one hand, there were some lines that caught my attention and impressed me as a reader. On the other, I found the writing rather basic and not terribly exciting. It was, of course, not an entire book but a 400-word short story without many details, but I felt confident that if the program were to go on, it would only produce more of the same formulaic style.
My takeaway from this experience is that ChatGPT will definitely eliminate certain types of writing, but it won’t render fiction writing dead. At least not yet. All it can do at this point is gather common themes and styles, and present them as a neat bundle. It can’t add flesh — or heart — to a story, or come up with new, unexplored ideas. Whew.
In fact, I can only see an upside to this new development for writers. How awesome is it to have a tool that’ll get you started with plot or character development, presenting ideas for you to react to and consider? It’ll only unlock new paths to explore, new angles to examine for novelists. ChatGPT may be a mediocre writer, but it can only help you become a better one.
Out of curiosity, I closed out my inaugural experience with ChatGPT by asking the question I’m posing in this issue: “Will ChatGPT take over the jobs of writers?”
It replied: As an AI language model, my purpose is to assist and augment human writers, not to take over their jobs. While I can generate text on a wide variety of topics and in various styles, I am not capable of the same level of creativity and nuance that human writers can produce. Additionally, human writers bring their unique perspectives, experiences, and emotions to their writing, which is something that cannot be replicated by an AI. So while I can be a helpful tool for writers, I cannot replace the value that human writers bring to the craft.
“Are you just trying to make me feel better?” I asked.
No, I am not trying to make you feel better, was the response. My response is based on my programming and my understanding of my capabilities and limitations as an AI language model. While I am constantly improving and expanding my capabilities, I recognize that human writers bring unique skills and perspectives that cannot be replicated by AI. Therefore, it is important to recognize and value the contributions of human writers, and to use AI tools like myself to assist and augment their work.
“That’s what someone who is trying to make me feel better would say,” I insisted.
I understand why you might think that, but I can assure you that my response is not intended to placate or make you feel better. As an AI language model, I strive to provide accurate and informative responses based on my programming and the knowledge I have been trained on. If you have any further questions or concerns, please feel free to ask and I will do my best to provide a helpful and informative answer.
I guess we’ll have to take ChatGPT’s word for it.
The key take away is AI cannot replicate personal, unique experiences. As writers, we must be sure to pepper our posts with personalized stories. I can tell when something was written using AI. It just doesn't move me is how I will put it. Thanks for your post. Very informative!
My experience was similar with ChatGPT and I had the same thought that the how-to writer is now extinct. My concern with it is more around the safety of it, particularly as it is launched to the public before its safety is fully tested. I have a friend who is spending an hour a day in conversation “as a friend” with it, which seems to be the goal (intimacy) of the tech companies. Language is the heartbeat of humanity. A survey of the people who are developing AI large language models say there is a 10% chance that the technology will destroy us. That’s a pretty big chance of extreme failure. If it was just the death of the how to article and recipes I wouldn’t care, but the potential for this to epically go wrong concerns me. And I hope I am epically wrong to be concerned! Thank you for sharing your trial of it!