I agree with most of my peers that using AI for content creation is an opportunity and a threat. We can’t deny that AI for content writing can research across the internet and synthesize information faster than our human minds can. It can also spit out content faster than any human writer. And both can be perceived as a threat. But AI used for content writing also allows writers to do more in a smaller amount of time, which leads to more time for working on other parts of their business (like writing this blog, in my case!).
Therefore, my initial loathing for AI has been replaced by a begrudging working relationship. The best content is not created by AI alone but with the invaluable and essential intervention of human writers and editors with the author’s best interests at heart. In a nutshell, this is the human element that AI for content writing cannot yet deliver.
A recent study researched the perceptions of professional editors and proofreaders toward the use of AI tools. The editors and proofreaders agreed that AI for content writing was useful to hasten the editing process and help with ideation. However, they also noted AI’s downside, which was its inaccuracy, inability to truly understand the context of the writing, and the amount of editing required to make the content piece acceptable online. In other words, AI creates the need for writers and editors.
It’s interesting that the research confirmed my own experiences. If you are out there creating AI content, please do so in tandem with an experienced human writer or editor. AI-generated content requires editing to avoid a low-quality, bland, disconnected, and often inaccurate final product. Here are the top three reasons why, confirmed by research!
REASON ONE: Posting Unedited AI-Generated Content Can Lead to Big Problems.
First of all, if you haven’t already heard, Google is cracking down on indexing low-quality, AI-generated content. Therefore, posting pure, AI-generated material could plummet your SEO. At best, your AI-generated content should be a very, very, rough draft of the final. Plagiarism is an ever-growing issue, and you don’t want your content to look and feel like everyone else’s. Never create an AI article and hit “post” without giving it an excellent, thoughtful edit.
My warning may sound like a no-brainer, but I understand that editing can be time-consuming and tedious. I also understand that one of the dangers of using AI for content writing is seeing those glorious words appear onscreen in the blink of an eye and the temptation to copy, paste, post, and go on to the next project before giving it another look. We can make some unfortunate decisions when we’re pressed for time, but we should never skip the editing process, no matter who produced the content. AI-generated content should always be reviewed and/or rewritten by humans. This is especially true when you recognize the weak spots of AI (yes, there are many!).
Recently, I was reading an interview transcript from a podcast. It identified each paragraph of text as spoken by the host or the guest. I’m quite confident that AI generated the script because, in several cases, text paragraphs were appropriated to the incorrect speaker. For example, I saw two paragraphs in a row that had been delivered by the host incorrectly ascribed to the host (first paragraph) and guest (second paragraph). I don’t know enough about AI to understand what happened, but I believe the transcript wasn’t edited or even proofread well before someone posted it.
A recent Wall Street Journal article tells the story of content writers who lost their jobs to AI, only to have the client call them up months later and hire them back. The reason? The AI-generated content submitted by the marketing company needed major edits that required too much work on the client side. While AI content writing may generate content quickly, only humans can edit it with the precision, tone, and client voice required to fulfill the mission of each content piece.
Retain the quality of your content. Treat AI as your intern, not your senior copywriter.
REASON TWO: In Editing, AI Fails to Recognize Context.
As confirmed in the research, I have found that AI doesn’t process the context of sentences well. Sometimes, AI just doesn’t “get it.”
AI proofreaders, like Grammarly, typically predict the next word of the sentence, based on a calculated probability. While AI works on the “word” level, a human editor reads a paper holistically to understand how each sentence contributes to the main points, even when reading a section or sentence.
For example, I wrote a piece on “reasons to hire a content writer” in the form of laments. One was, “I need a new website!” and another was, “I can’t keep up with social media!” Grammarly didn’t understand my format. It wanted to change the third lament from “Nobody understands what our company does!” to “We need to understand what our company does.” The recommendation changes the point of view and tone of the messaging from a CEO’s lament about their current confusing website to a need for a mission statement. The distinction is significant; again, AI content writing just doesn’t “get it.”
Also, unless you input a great deal of specific information into AI, it does not consider the author’s preferred tone and word choice. I do a lot of work with authors who are writing personal essays and biographies, which is work that AI content writing cannot generate. My mission is to help the author deliver a grammatically correct piece but also preserve their voice as much as possible. When it’s time to have AI proofread a piece for me, it’s clear that the bot has no concern over retaining author’s voice or preserving their intended meaning. Proofreading tools aim to simplify and shorten writing as much as possible. Period.
I’ve repeatedly experienced how AI attempts to “over-edit” a piece when the author wants to project their style. I find I must review every recommended edit carefully for context. Anyone who thinks they can just “accept all changes,” will be disappointed and eventually need to spend a lot of time changing back edits when they see the final product.
It’s possible one day AI for content writing will be sophisticated enough to understand our intentions and preferences. For now, we must be careful never to accept blanket AI recommendations and consider how each one individually contributes to the overall messaging of the content piece as well as the author’s intended voice and preferences. If that sounds like a human writer or editor job, it is!
REASON THREE: AI Gets Creative….and Misrepresents.
We all fact-check anything AI gives us, right? You should know that sometimes AI misrepresents things and sometimes, it just lies! When AI doesn’t provide source citations, things can get very interesting.
I performed an AI experiment with me as the subject. I gave ChatGPT a prompt to write a biography of me. It’s an eye-opening exercise, and I highly recommend it. I thought the first draft was quite impressive and complimentary, although Grammarly suggested 15 edits which I have not implemented so you could see the raw data. (AI suggesting edits on AI.I guess nobody’s perfect.) I also found the bio to be accurate, but a repetitious, jargony yawn.
Karen Dix is a skilled content writer with extensive experience in crafting engaging and effective content for various industries. Over the years, she has developed a strong reputation for delivering high-quality copy that aligns with the specific goals of her clients, helping businesses communicate their messages clearly and persuasively.
Karen’s versatility as a writer allows her to work across multiple formats, from blogs and articles to web content and marketing materials. Her ability to adapt her tone and style to fit the audience has made her a valuable asset in both B2B and B2C environments. Her content is always well-researched and optimized for SEO, ensuring that her clients’ messages not only resonate but also rank well in search engines.
Her background in content strategy and digital marketing has equipped her with the knowledge to create content that drives traffic, builds brand authority, and converts leads into customers. Whether working with startups or established brands, Karen has consistently demonstrated a commitment to helping businesses grow through the power of content.
Next, I asked it to make the bio “more fun.” It was a valiant effort, and AI certainly fulfilled the mission. However, the underlined portions made me wonder where AI got its information. AI got a little bit “creative” to the point of misrepresenting me a bit.
Karen Dix is a wordsmith extraordinaire who turns ideas into captivating stories and dull copy into pure magic. With a keyboard in hand and a coffee cup nearby, she’s been helping businesses of all sizes connect with their audiences through content that not only informs but also entertains.
Known for her quirky sense of humor and a knack for making even the most technical topics relatable, Karen’s writing has found its way into blogs, websites, and marketing campaigns across industries. Her secret sauce? A blend of SEO wizardry, creativity, and the ability to speak to readers like an old friend. Whether she’s crafting snappy headlines or in-depth articles, Karen knows how to grab attention and keep it.
When she’s not busy writing up a storm, you can find Karen immersed in a good book, exploring new podcasts, or perfecting her home-brewed coffee game. She’s passionate about storytelling and has a soft spot for helping businesses tell theirs in a way that sticks. Fun fact: She’s been known to use puns in her copy—strategically, of course.
AI thinks it knows me. It guessed at the “coffee cup nearby” detail, and sure, it was correct. I am a coffee drinker. If I drank tea, it would work as well, or if I drank cola out of a coffee cup for that matter. I suspect most writers fuel themselves with coffee or tea, so AI made a good guess. However, I’m not sure my clients would identify me as someone with a “quirky sense of humor.” I also don’t think I use many puns in my copy (another safe assumption for writers), although that’s up for debate. The blatant falsehood here is the home-brewed coffee game. It sounds fun but nothing like the games I like to play. As usual, AI for content writing produces general, non-specific, ingenuine content that is a fair first draft but can definitely be improved upon by a human writer or editor.
I recently attended a seminar on AI in the workplace, which addressed the importance of rolling AI into the workflow process as the first step on every project. With more specific prompts, you can see how AI gets better (and I’ll admit it, a lot more fun!), but you can also see how things can get crazy very quickly.
The tech world calls these fabricated truths “hallucinations,” which is the fake stuff AI for content writing creates that will nonetheless continue to be regenerated all over the internet as truth. And since I found some in my AI-generated bio, I thought I would try the same experiment with some other people I know. I found that AI fabricates lots of details.
- FACTS. I caught AI saying my friend attended high school in the town they currently reside in. In truth, my friend has lived in the town for twenty years, but was born, raised, and attended high school in another state.
- INTERESTS. AI described one of my introverted, homebody friends who lives in Chicago as someone who can be found soaking up the city’s art and music scene. (Not!)
- EXPERIENCE. I credited a local writer friend with having been published in national publications. She’d be thrilled! A student I know who has not graduated college has already established himself as a leader in developing cutting-edge systems in his field!
- AUTHENTICITY. From my experience, AI likes to take a stab at interests, based on profiling the person and making general assumptions. AI can make probable guesses that aren’t always accurate, which is another reason human editing (and sometimes major editing) is needed.
This is just a small, partial list of hallucinations by AI, specific to my experiment to make a point. As humans, we should always be sure that what we are reading is reliable and from a credible source. When we create content, we have the responsibility to be that credible source and ensure that everything we send out into the world is, indeed, accurate and honest.
After I read the AI-generated bios, I wondered why AI took a guess about the coffee game in mine when it could have omitted the entire detail. I guess that’s where hallucinations kick in. AI is just kicking back and opening its mind…like the hippies did in the sixties. AI is brave, bold, and tries to be inventive. So again, make sure a human writer or editor fact-checks everything in AI-generated content and makes other necessary edits too.
Speed vs. Quality
While AI will always beat mere mortals in the speed of creating content, it is still up to human writers and editors to be the gatekeepers of quality. Only a real live human can whip an AI-generated content piece into one that will engage the reader, delight the client, and perform in search engines as needed. Remember, treat AI like you would an adversary in a tough negotiation. Never accept a first offer. Rewrite, edit, and make your content sing!
I am a real live writer and editor bent on producing the highest quality content for my clients, with or without the help of AI for content writing. Let’s discuss your content needs and the best strategy for you. Contact Karen at karen@bigideaswriting.com or 630-778-6182.
Leave A Comment