ChatGPT. It’s all anyone can talk about. At this point, we all know what it is, but for copywriters and marketers, how do you use it?
At Scribewise, we’ve tested different uses for ChatGPT to see how it will fit into our lives and our jobs going forward. Here’s how we’ve experimented with it and what it’s taught us.
Tested its ability to plagiarize itself
Since the ChatGPT model is trained on a large dataset, not the internet itself, it’s incapable of plagiarizing published or written work by humans. But we wanted to know if it could plagiarize itself. If two separate people asked it to write about the same exact concept at the same exact point in time, how different would the results be? Will they be distinguishable from each other?
Prompt: What is insider risk?
User Response 1: Insider risk refers to the potential harm that can be caused by employees or other individuals within an organization who have access to sensitive or confidential information. This can include things like the theft of sensitive data, the intentional or unintentional sharing of confidential information, or the sabotage of systems or networks. Insider risk can be difficult to detect and prevent, as it often involves individuals who are trusted members of an organization and who have legitimate access to the information in question. Because of this, it is important for organizations to have policies and procedures in place to identify and mitigate insider risk, and to regularly monitor and review their systems and networks for potential threats.
User Response 2: Insider risk refers to the potential for harm that can be caused by individuals within an organization who have access to sensitive information or materials. This could include employees, contractors, or other insiders who may intentionally or unintentionally cause harm to an organization by leaking confidential information, stealing data or resources, or sabotaging systems or networks. Insider risk is a growing concern for many organizations, as the potential for damage from insider threats can be significant.
The short answer is, basically, no. While the results were worded somewhat differently, the content was the same. Without more direction, the chatbot will use a similar voice and similar words to deliver your results every time.
Examined its copyediting capabilities
As we’ve now seen and tested out, the chatbot’s writing capabilities, while decent and sometimes clever, are pretty uninspired. But why not use it as a tool for copyediting?
After feeding ChatGPT a couple of blogs, I was surprised by its ability to maintain my writing style and tone of voice. Most importantly, it rarely cut out important information. Rather, it just cut out fluff, summarized my points more succinctly, and shifted some words around. Given as much context and detail as possible, it could also keep my blog to a certain word count, apply certain grammar rules, and even summarize my long prompt into an outline if asked.
Of course, these are things I could have done given an extra half hour or so. But why spend the extra time turning your first, messy draft into a cleaner, second draft when a bot can do it in a few seconds? While I still had to proofread ChatGPT’s new draft of my work, doing this helped me observe where I could improve my own writing while saving time for me—and my copyeditor.
Asked it to teach us something
Almost everyone (even Google) agrees that one of ChatGPT’s best use cases is as a search engine. As writers for various and often complex industries, we spend a lot of time learning and researching subject matter before we can write about it. We’ve found that ChatGPT is good at cutting through the noise of ad-riddled search engines pretty well.
For example, while trying to write an article about how mirror therapy works as a physical therapy treatment for amputees, I first needed to understand how neural plasticity works in the brain. So ChatGPT, what is neural plasticity? And then, how does it impact amputee patients in physical therapy? And then, how does it work with mirror therapy?
Of course, ChatGPT isn’t always right—sometimes it just makes stuff up. But luckily, with the help of peer-reviewed research and a conversation with a subject matter expert, I could actually fact-check this information before allowing it to inform my writing.
We all laughed when users first tried to break the chatbot because we realized ChatGPT has a sense of humor. (Sometimes, when their site is so overwhelmed that no one can use it, it gives you an error message in the form of an acrostic poem or 90s hip-hop rap lyrics, explaining why you might have to be patient while it’s “at capacity.”)
But we had our fun too. Why not ask it to write a poem about pizza while you do an aloud reading to your colleagues, we thought?
Pizza, oh pizza, with your intoxicating smell
A mixture of dough and sauce, a symphony as well
Your toppings, oh your toppings, they make our mouths water
Your smell is a siren’s call, it makes us want to slaughter
All other food, for we know that nothing else can compare
To the mouth-watering aroma that fills the air
When we smell pizza, our senses come alive
We can’t help but salivate, and our hunger does thrive
So here’s to you, pizza, with your irresistible smell
A scent that can make even the grumpiest person well
We love you, pizza, and your delicious aroma
For you are the king of smells, and we bow down to ya.
So, what have we learned?
This thing is not going away, so as long as it’s part of our future, we may as well become experts in it, right?
Overall, we agree with digital marketer and bestselling author Ann Handley: “We writers can’t passively sit back and let AI write *for us.* The way to use AI is as a gymnast using a spotter and a coach—a way to help you create with more confidence. Even fearlessly. Yet it’s your talent that drives AI. You are the gymnast!”
Use this chatbot to make you write better, not harder. There’s no question that writing is hard work, but if having your own personal copyeditor, idea generator, summarizer or intern in your back pocket helps you write more, faster and better … Why not try it?