One of the more interesting (but no less terrifying) developments in 2023 has been the popularity of artificial intelligence (AI) writing tools. Some options are available free online, such as ChatGPT, which also has a paid Plus plan. Others are only accessible through a subscription (Jasper, for example), and some of these primarily target businesses.
The fact that many people say this technology will replace me at my job someday only made me more curious to learn about it. (C’mon… replace ME? Who’s going to drink all the coffee at the office?)
If you’re a writer considering delving into AI—or wondering if it might be time to learn a new trade—this blog is for you.
For over a month now, I’ve had access to an AI tool specially developed for content creation. Its intended purpose is to help me expedite the creation of all types of marketing copy and content—ads, blogs, press releases, social media posts, emails, and more. Note how I said the AI is there “to help me”—let’s not ever forget that, okay?
If you’re going to work with AI, there are a few things you should know. Something my colleagues and I learned early on was that AI has no trouble just making stuff up, particularly when it comes to citing its sources. It always provided authentic-looking URLs as references, but many were not valid.
In another instance, I was using it to help me construct a list of selling points for a well-known product. Periodically, it would inject language like, “Try it free for 60 days”—an offer that my research (and Google) could not corroborate. (Apparently, AI has a little cyber used-car salesman built into it, so always check its “facts.”)
Speaking of research, AI also has a few biases that would taint its findings, depending on your subject. For instance, it’s not particularly fond of humans and believes them to be a threat to the planet—but that’s all. (Nothing to see here.)
Then, I observed that AI never expresses a lack of confidence in its responses, except to imply that a particular outcome may be uncertain. When I asked it a paradoxical question, instead of trying to answer, it properly identified the question as a paradox and even told me why.
Honestly, I found that response refreshing versus trying to unravel the paradox. It was like artificial humility.
From those experiences, I learned the importance of asking AI the right questions, which in this context are called “prompts.” This is a skill unto itself and recently gave birth to the phrase “prompt engineering.” It takes some practice to get familiar with AI’s strengths and quirks. In general, the more specific you are with your prompts, the more successful the AI is at generating the output you desire.
So, for example, if I prompt AI with “Write 3 paragraphs about digital marketing trends,” the AI will output very high-level (non-specific) trends, like mobile technology, social media, and (ironically) the potential of AI. (Did I mention that AI is openly narcissistic?)
But—if you wanted to write a piece that focused more on specific technologies and how they might shape digital marketing trends, you’d have to bring that detail into your prompt, like this:
“Write 3 paragraphs about how digital marketing trends have changed in the past 2 years and how they might change in the near future. Focus on technology.”
From this, you’d get more detailed future-looking output that might include using strategies like personalization and segmentation along with specific AI-based marketing tools and perhaps the adoption of voice user interfaces (VUI) and augmented reality (AR) technologies.
Something that is really powerful is AI’s ability to keep things in context as you input new prompts. For instance, in the previous example, if you wanted to add another paragraph that focused on the role of AR in digital marketing, you’d need only input a new prompt like: “Focus on augmented reality.” The new output will automatically be in the context of the previous conversation, which in this case was about the future of digital marketing.
That’s really helpful when creating long-form content because you can build it out one piece at a time, like bricks in a wall. You don’t have to keep reentering a whole new prompt until you get the desired output. That gives AI a nice advantage over the old-school web-browser search engines.
Overall, the best advice I can give when communicating with AI is to ask questions in the same way you might describe a painting to a friend. Be very cognizant of what you include—as well as what you exclude—to avoid diluting the output. To AI, everything in the prompt tends to be equally important. (Except people, of course.)
In this blog, I’ve shared my first experiences as a copywriter collaborating with AI. This is just the beginning of the journey, and it will no doubt get more interesting as I get better with the technology and the AI gets “better” as well. I look forward to sharing and revisiting the million-dollar question: Will AI replace me? (For now, I think not.)
As a writing companion, AI is great for conquering the blank page and quickly assembling a half-decent first draft. But for now, at least, it lacks those human qualities that connect us to a good story or succinctly build a convincing case.
As with most joint efforts, the end results are better by virtue of the collaboration itself. That is, I believe the copy I produce using AI is better than what I might create on my own, but at the same time, I’m making the AI’s output better as well, so… touché.
For the long-form content I typically produce, collaborating with AI helps lay the foundation more quickly so I can focus on the more nuanced facets of word play. It provides different points of view to consider, which allows me to cherry-pick what I like best to create the most impactful copy.
If you are a copywriter who has been on the fence about incorporating AI into your creative processes, I hope my experience encourages you to test the waters.
And no AI was used in the writing of this blog. (This time.)