Should You Use ChatGPT & Other AI in Your Marketing?

Imagine typing a short prompt into a free online machine, and within minutes it generates a 1,000-word, well-written article as requested. All with perfectly-written sentences, no grammatical errors, and exceptional in-depth research. ChatGPT allows you to do just that.

There’s a new fad in town; at first, it was pretty scary news for me, a lifelong marketer who makes a living writing copy, creating content, managing social media for clients, etc. So, upon hearing about the public launch of ChatGPT, I had to dive deeper into AI-created content and find out just what happens if you rely on this new tech for content and marketing.

Like many of my peers, I ran – not walked – to my computer to try my hand at ChatGPT. I typed in a new chat query and watched in awe as an essay scrawled across my screen in seconds. I was so shocked that I even texted a screen recording to my husband. My initial thought was: Oh shit. This could be really bad for me. 

What is ChatGPT?

In 2015, a group of Silicon Valley investors, including Elon Musk, created an AI research non-profit, Open AI. The goal was to train the AI for conversational engagements like customer service chats. On November 30, 2022, Open AI launched ChatGPT to the public, which has since disrupted the industry.

ChatGPT is a technology that takes basic instructions and creates responses using learned information. It understands human speech and creates in-depth writing by scanning the internet and using syntax similar to human-to-human interactions. In other words, you can ask ChatGPT to write an essay on anything, and it will create a completely AI-generated essay or blog for you.

My Experience with ChatGPT

I am a natural doubter, so I obviously have reservations about what sounds too good to be true. 

My first thought was to test ChatGPT for myself. After running a query and receiving the finished content, I would take the article and run it through editing software, Grammarly, to check for plagiarism and get an overall grade of the content. 

To be transparent: I run all of my copywriting through Grammarly before submitting it to clients. It is an everyday tool that helps me check my work for grammar, correctness, and plagiarism. Grammarly provides me with a content rating out of 100. For comparison, my typical human-written articles start with a composition grade of 97 percent and end with a grade of 100 percent. My average plagiarism grade is within 1-2%, which is extremely low for researched content. It’s important to note that most journals and news outlets widely accept less than 15% text similarity.

I asked ChatGPT to write an article using the following query: “Write an essay on England's royal family.” ChatGPT generated a 380-word essay on the British monarchy within one minute. 

I then copied and pasted the text into my handy-dandy tool, Grammarly. It yielded a composition grade of 85%, a solid B. Grammarly also detected plagiarism at 6%. By conventional standards, this is nothing to raise a red flag. 

Grammarly rates the copy by identifying any writing issues. The ChatGPT-generated article was flagged for clarity, correctness, delivery, and engagement. In addition, Grammarly also noted the text used fewer unique words than usual. It ranked the overall readability of the essay using the Flesch reading-ease test at 43. In comparison, this essay reads like a NY Times article and is likely to be understood by anyone with at least a 10th-grade education.

After running the ChatGPT-generated essay through Grammarly and making the suggested edits for grammar and conciseness, the text scored a 100%. 

Pretty good for ChatGPT. Pretty bad for me, I thought.

I then started to look into the consequences of using ChatGPT for my line of work. Is AI-generated text detectable? Can website owners be penalized for taking the easy route? Will using software like ChatGPT hurt me and my marketing? Here’s what I found out…

Should You Use AI for SEO & Content?

First, if it looks like a duck and quacks like a duck, it probably is… a duck. AI-generated content is no different. 

Is AI Content Detectable?

Technology like ChatGPT is creating quite a stir in academia, and it’s obvious why lazy students may be drawn to the software. But is AI content detectable? There are currently two tools to detect AI content. One is GPTZero, an app created by Princeton student Edward Tian. The tool checks the text for perplexity and burstiness. In other words, the frequency of words used and the syntax of each sentence. A human is likelier to use varying sentence structures, whereas AI-generated sentences are more uniform. This is reflected similarly in the Grammarly results, where I found fewer unique words than in a typical text.

The second tool used to detect AI content is provided by the creator of ChatGPT, Open AI. GPT-2 Output Detector is a simple copy-and-paste tool that allows you to drop text into an empty field. It then predicts if a human or a bot created the text. I had to test this for myself, so I dropped the Grammarly-edited ChatGPT-generated content into the detector. It ranked the content as 99.98% Fake (or written by an AI machine) despite being edited. 

This test proved my point: I will never use AI-generated content for my clients. It is detectable and likely to be biased or flagged for misinformation.

Open AI has also released a statement saying they will watermark ChatGPT-generated content, making it easy to detect the use of AI technology in copy.

Will Google Know if I Use AI Content? Will Google Penalize Me?

John Mueller, Google’s Search Advocate, announced in April 2022 during a Google SEO Office Hours interview that AI-generated content violates Google’s guidelines and is considered spam. Since the dawn of Google, the company has actively worked to police and monitor AI-driven content, and the lines are clear. Mueller explained that the Google webspam team will take action if they detect AI-generated content on a site. He does specify that using AI-generated content in any capacity is considered spam. 

“Currently it’s all against the webmaster guidelines. So from our point of view, if we were to run across something like that, if the webspam team were to see it, they would see it as spam.” - John Mueller, Google Search Advocate

You can watch the full interview below. I’ll warn you in advance, it’s 57 minutes long and miserably dull.

My Takeaway

In my opinion, ChatGPT generated a well-written, well-researched article in the fastest turnaround time. Does that mean I’ll be using it in my day-to-day work as a marketing expert? Probably not.

My sole job is to protect the reputation of my clients online. I guard my clients’ Google ranking like a mama lion, so taking a shortcut falls flat when the repercussions are clear and defined. AI-generated content is excellent for quick research. For example, if I need fast facts and don’t have time to sift through Google results for my answer, ChatGPT may help me get there faster. But I don't see myself leaning on AI bots when crafting a brand voice and authoritative content for my clients. 

I believe detecting AI content will become easier in the near future, and with these detection tools comes stricter enforcement. I will fiercely warn my clients against publishing AI-generated content, the same way that I advise against buying likes or followers on social media. The shortcut is always appealing, but the fruit is in the long game. 

Are you as interested in this topic as I am? Continue the conversation by email at hello@yasminparsloe.com.