In the past 12 months, searches for “how to humanize AI content” have increased by 943%.
More and more people are trying to “pass human censorship,” either making AI content less bad or touting AI content detection tools.
The process of “humanizing content with artificial intelligence” has become the de facto content strategy for companies eager to grow.
I see pent-up writers popping up everywhere looking for advice on how best to tackle this headache.
I know how soul-destroying this is because I own this t-shirt.
That’s why it was doomed to be a failed endeavor from the start.
We all know that there are obviously big problems with artificial intelligence content (if you don’t know, you can read about these problems) here), but I don’t think anthropomorphizing robots is the solution.
In fact, I think humanizing long-form content generated by artificial intelligence is an inherently flawed process.
The whole point of using AI is to increase speed, scale, and efficiency; having humans edit or even rewrite long-form AI prose completely undermines this value proposition.
As you research and add a “rich” human element, you may have created the content from scratch yourself.
and through a process of running “humanized” AI content detector This is ridiculous – you’re essentially asking artificial intelligence to validate humans.
I would spend hours editing the AI content I was responsible for so that it would pass human review, only to be fired up and start thinking about my content was Artificial intelligence…artificial intelligence?
And I’m not the only victim. in a LinkedIn PostKiran Shahid, a content consultant, said AI detection tools only validated her humanity when she wrote in the first person — a writing style that clients were less likely to recognize.
I saw similar experiences online Women in the Tech SEO Community…
continue Reddit…
Having human writers glorify AI-produced content, only to have AI evaluate their humanity is not only bizarre, but a blatant waste of talent.
When I’m asked to “humanize artificial intelligence,” my workflow looks like this:
- Receive batches of 100% AI content for editing
- Ensure each article is at least 50% “human” in AI detection tools
- Turn all circles in the SEO Traffic Light tool green (Pro tip: keyword content. Apparently it will work wonders in 2024!)
The AI articles sent to me were so out of date that I essentially rewrote them (but with the added pressure of unrealistic deadlines), and even then the end result was subpar.
When you shoehorn in examples, ideas, and insights that could affect the content of the article if they were included from the beginning, the flow of content will inevitably suffer.
At this point, I would like to clarify something. I’m not against artificial intelligence. I use it every day to ideate, structure, outline, and repurpose—my number one workflow is improving the pacing and readability of what I write in “the process.” Used like this, artificial intelligence can be very efficient.
But the process of “humanizing AI” — starting with 100% AI-generated content and then animating it, like Mary Shelley breathing life into Frankenstein’s monster — is one of the most trivial of tasks. One, and not only does it demotivate writers (although that’s cool I just said it).
If you’re someone who implements this strategy, you should also know that it’s a great way to churn out talent.
Humanizing AI content is unsustainable and inefficient, but some companies will let you do it anyway.
If you’re in this situation, one way to make your life a little easier is to train the AI on your content from the start.
This means that every time you receive a batch of AI-generated articles to “humanize,” you can run them through a self-trained AI model and bring them to a more accessible level of humanization. Then go back and make your own edits and additions.
Training an AI model is a smart move because it improves the AI system itself; otherwise, it’s just laborious manual work and you might as well have written the content in the first place.
Tools like Claude allow paying users to create shared “project” contains up to 200,000 words of training documentation, as well as Custom GPT for ChatGPT Allows you to train up to 20 files with a maximum size of 512 MB each.
Nishkasha KotianSEO expert channel engine Provide examples of her brand guidelines to a custom GPT and create rules for the bot to follow.
Similarly, Melissa PopeDirector of Content Strategy The shaky Luconduct knowledge document training on her customized GPT, and adopt “Tips for a few shots“By providing example Her best content.
Since AI tools have “memory”, you can also practice “iterative refinement” while chatting with them.
In other words, you can train them bit by bit using suggestions and examples to improve the overall content output. Digital PR Expert Gini Dietrich describes artificial intelligence In this regard, as a “digital intern”.
The problem is, training takes up a lot of time, and even with all the background in the world, if you don’t have a clear idea of what your digital interns need, you can easily get stuck in an endless loop of prompts.
So training an AI model can take some of the burden off, but it’s far from foolproof.
Melissa Pope Pointing out that customized GPT only allowed her to achieve 40-50% of her goals, and the effort spent on training often undermined its benefits.
Questioning “efficiency” strategies like humanized AI can be tricky—especially when they are in high demand.
Decision makers often don’t understand the reality of the job and are only concerned with turning the numbers in a spreadsheet green.
But your job is to make this reality impossible to ignore.
In my case I took the following steps:
- Regularly provide feedback in meetings on the problems I encountered in the “humanization” process
- Tracking the time I spend “humanizing” artificial intelligence in collaborative spreadsheets to demonstrate time consumption
- Outlines the risks of launching sub-par AI content and provides examples of brands that have struggled in search results pages due to manual or algorithmic penalties
One approach I haven’t researched but you could try is to direct stakeholders to AI content that doesn’t require too much human touch, such as:
- Glossary
- Product Description
- FAQ
- Wikipedia
If the end goal is traffic, you might also consider coming up with non-AI initiatives that have been proven to drive traffic at scale.
In a recent study by Ahrefs 50 bootstrapped SaaS companies will dominate SEO by 2024the brands driving the largest year-over-year traffic growth have adopted three core strategies:
- “Free tool” creation
- stylized content
- Content localization
Non-AI projects don’t have to be time-consuming either. ZapierFor example, thousands of feature-based landing pages were programmatically built and then outsourced to partners during onboarding.
This kind of traffic play is more likely to meet the rapidly growing needs of customers/executives, leaving you with time to focus on the work of creating quality content.
final thoughts
Humanizing content generated by artificial intelligence is unsustainable and, contrary to what many people think, inefficient.
AI should be used for ideation, planning and drafting from the start, not ghostwriting.
in speech Senior Content Marketing Manager at Moz, Chima Mmeje:
AI content is not the problem. The thing is, when the content is so poorly written, I can tell it’s artificial intelligence generated. AI detectors are garbage. If you hire a writer and you put your content through a detector as evidence of its appeal or authority, you should not work with the writer. Writers, I implore you to stop using artificial intelligence to create your entire content. Please use it as a guide to help you flesh out your ideas, work through your thought process, and even clean up your writing, but the first draft should be 100% your work.
If you’re reading this and you’re capable of strategizing, give up on the soul-destroying process of humanizing AI content.
Believe in your humanity and write content for humans.
