I rarely disclose my secrets of the trade (except for some of the advice I share on TikTok). But when it comes to AI and content creation, this is one area of my craft that I’ve been experimenting with and honing for several years, and I must share what works.
I am talking about how to humanize AI content. Platforms like ChatGPT, Perplexity, and Jasper have become mainstream in every writer’s toolkit. But they also come with quality control concerns, such as a high percentage of text with patterns resembling AI-generated content.
Depending on who or what you’re writing for, content dense with AI-written text is a valid concern. The agency you’re working for might scoff at your article that has 55% detected AI, wondering why they hired you when, presumably, all you’re doing is using AI for the heavy lifting.
That’s where humanizing AI content is crucial. Because let’s be real, the efficiency (and even the quality) of work is hard to beat when using AI mindfully. Before AI, it would take me around 6 hours to write a high-quality, 2,000-word article (depending on the topic). With the help of AI, I’ve cut that time in half.
To achieve this level of efficiency without compromising the quality of my work, the human element remains critical. Below, I share some of my strategies and best practices for humanizing AI content.
How to humanize AI content: 5 strategies that work
Enough introduction. Let’s discuss the strategies and techniques I use to humanize AI text. The beauty in these strategies is that you can apply them immediately (and see immediate results without investing hours into human editing).
1. Include technical details in your prompts
In my prompts, I have a dedicated section for “technical rules” that lists things I want the AI to avoid. These are common patterns that AI loves, but humans rarely use. This is a direct copy and paste from many of the prompt templates I use:
“Technical rules: when writing content, please avoid:
- Lengthy paragraphs over three sentences
- Using em dashes (“—”)
- Comma-linked clauses (or comma verb-ing modifiers), like “Missteps can lead to X, making Y a strategic priority.”
- Filler phrases that say nothing, like “At its core,” “In the ever-evolving landscape of,” or “It is crucial to recognize.”
- Overused comparisons that sound robotic, like “More than just a solution, X is a” or “Not just a tool, but a.”
- Unnecessary poetic language, like “A tapestry of creative ideas” or “A symphony of insights.”
- Overuse of the “x, y, and z” structure, like “AI improves efficiency, boosts productivity, and enhances workflow” or “This tool is fast, reliable, and easy to use.”
- Repetitive introductory sentences, like “When it comes to X, it’s essential to consider Y” or “A key factor in X is Y.”
Feel free to copy and paste this list of technical rules and watch as your AI-generated content sounds dramatically more human by avoiding these patterns.
2. Harness the power of writer personas
Instructing AI to write like a particular person is a powerful way to get more human-like outcomes.
In addition to technical rules, I include a section that literally tells AI to follow a specific writing persona, such as Seth Godin, and tailor the content to sound less like a sales pitch and more like a deeply human, lightly investigative story.
I’ll even include an example of Seth Godin’s work and ask the AI to mimic the model I am feeding it:
“Great stories succeed because they are able to capture the imagination of large or important audiences.
A great story is true. Not necessarily because it’s factual, but because it’s consistent and authentic. Consumers are too good at sniffing out inconsistencies for a marketer to get away with a story that’s just slapped on.
Great stories make a promise. They promise fun, safety or a shortcut. The promise needs to be bold and audacious. It’s either exceptional or it’s not worth listening to.”
Find an author, blogger, writer, persona, etc., that closely resembles the human voice you’re trying to achieve, and incorporate these instructions and examples in your prompts.
3. Personalize your AI content
Eradicate the AI in your text by taking things personally.
One of the best ways to humanize AI content is to use a first-person writing perspective using words like “I,” “me,” and “my.” In other words, personalize your content and speak from your own original voice.
I realize this is not always practical for certain clients or applications of your content. Oftentimes, we’re working for other businesses and brands under strict writing guidelines. But this is a highly effective way to reduce AI-detected content and better engage your readers.
While there isn’t any hard proof that personalized content is better for SEO, my theory is that search engines also see greater authenticity and value in personalized content.
If you’ve heard of EEAT (Experience, Expertise, Authoritativeness, and Trustworthiness), then you’ll get what I am talking about. This leads me into my next strategy for humanizing AI-generated text…
4. Leverage human expertise and experience
In parallel with personalizing content, you can further reduce the artificiality of your work by incorporating more human expertise and experience.
This doesn’t have to be limited to your own expertise and experience. Because after all, seldom are today’s content writers writing for themselves. But you can include original quotes and personal anecdotes from other subject matter experts.
One of the agencies I do freelance writing for has implemented a rule requiring at least 1 to 2 expert quotes in all blog content we produce. This sort of editorial approach is quite common in high-performing content and successful blogs.
In addition to quotes, consider incorporating examples, experiences, or use cases that resonate with specific audience personas.
For instance, if you’re writing a cybersecurity article about AI deep fakes, you could share real-world examples and case studies that highlight the magnitude of these threats. Further, you could include solutions geared toward business owners, high-profile executives, or celebrities who are most affected by deep fakes.
Zooming out, your focus should be on addressing real human experiences and validating your work with real human expertise.
5. Utilize AI as a researcher and co-writer
I treat AI as my research assistant and co-writer, not my ghostwriter. Instead of asking it to generate full articles in one prompt, I request individual sections, bullet points, relevant data, or key statistics. This gives me the raw materials I need to expand and write in my own voice.
For example, when I’m tackling a comprehensive guide on a technical topic, I’ll ask AI to gather data and statistics for me. I also use AI summarizing tools to pull out key points from articles, reports, and even audio or video files. I take that research, analyze it through my own lens, and articulate it in my unique writing style.
Think of AI as your collaborator. It can accelerate your processes, but it’s not a substitute for human creativity and critical thinking. I’ve found this co-writer approach strikes the perfect balance. I get the efficiency boost without sacrificing the quality or authenticity that makes content actually worth reading.
Important takeaways
Humanizing AI content is NOT using tools to do the work for you. I cringe at how many AI-humanizing tools have surfaced in recent years. What’s even more cringeworthy are the outcomes these tools generate. They butcher your content to make it sound like a lifeless podcaster filling the air with clichés and wasted space.
Sure, things will likely change in the near future as AI advances. But for now, you can’t fight bad AI content with even worse AI content. In short, the tools used to humanize AI text are largely ineffective in my opinion, and I don’t suggest using them.
Hopefully, these strategies are useful to you. If you have any questions or want to connect, please don’t hesitate to hit me up.