The most significant challenge with AI writing is that most can recognize it at once—the slightly robotic, excessively sleek tone, the tone that could be a press release from someone who has never talked to a human. Learning to make ChatGPT content sound natural is essential for creating authentic, engaging content that resonates with readers.
The positive? If you have the correct editing skills and a true knowledge of your audience, you can adapt turgid AI material into authentic content through enhancing AI writing techniques.
First of all, here is the information you may be missing.
Why AI writing sounds robotic initially
ChatGPT is trained to be helpful, truthful, and harmless. Those are very good qualities. Yet they also create writing that appears to be too balanced, too formal, and strangely monotonous in its cadence. To have equal weight on each sentence. All the paragraphs manage to "mousetrap" themselves.
No extra mess, no personality, no feeling like it was actually written by a real person—someone with opinions and bad days and strong feelings about coffee.
Examples that appears in many answers for common issues are (among many other): - Repeating structure of the sentence - AI loves the subject-verb-object sequence. Again and again.
If we understand what we are dealing with, the following issues become clear: -Multiple use of transition words such as "Furthermore" and "In conclusion" -All of the text becoming oversaturated with brackets such as "it is important to note that…" -Indiscriminate use of stuffing language such as 'it is worth considering that…" -Indigestible generalizations rather than exciting and precise illustrations -Absence of emotional register - While it is all grammatically correct, it is dead boring!
Methods to make ChatGPT content sound natural
Use meandering sentence structures. Just breaking the rhythm can humanize text in a flash. Examine a paragraph of ChatGPT output and you will notice that each sentence is close to similar length—say, 15–20 words. It's a bit tiring to try to read all that even if all was interesting is something reliable.
For longer writing, make your sentences about half the length. Or even more concise. Then allow a single long sentence—allow a few concepts to go along for the ride, the way participants do when getting in-depth about a subject that matters to them.
Use contractions. It actually sounds a little too easy but it does make a difference. Hence, you will find that it is helpful becomes Hence, you'll find it's helpful. That's not just shorter - but warmer too. It feels more like someone is speaking to you instead of speaking at you.
Include some kind of the point of view. People see the neutrality of the AI writing very well and even may take it for granted. That's not a big deal when you're writing a Wikipedia article, but it's causing me a heart attack when I might use it as a blog post, email, or Tweet.
Feel free to put in opinion: "I think this method works better," or "from what I have observed, most begin this process and regret it." Opinions foster relation. They indicate human judgment.
Substitute "abstractions" with "specs". This is perhaps the most effective tactic. The AI is more likely to say something such as there are a number of methods that may lead to effective engagement whereas the human author will be more positive such as posting at 7 a.m.
Details seem to stick in your mind just a little more than any other kind of writing. "On Tuesdays tripled my comment rate—at least for that audience." Specific details are just a little more memorable. Vague overviews are not.
Common AI writing problems and solutions
Problem number 1: AI Version: "This method is very advantageous. It has also been proven to boost productivity. Furthermore, people are usually quite appreciative of it." Human version: "It has advantages. Productivity even rises—and strangely enough, staff seem to actually welcome the increase, which is more unusual than you might imagine." The solution is straightforward: eliminate the transitional throat-clearing and allow ideas to link together naturally.
Problem: hollow enthusiasm which is the AI he's uttering: 'This is a very powerful tool that is going to change the way that you work' Human version: 'This tool does indeed save some time - approximately an hour a day if you use it everyday' Truth still wins over hype.
Passive voice too much. AI version: The results have shown several factors influenced the outcomes. Edit version: It has been found that many factors caused inaccuracies in the results. (mainly sample size/timing.) Active voice is said to be faster and more comprehensible and sounds more human.
Tips for humanizing AI content effectively
READ it. In. Out loud. Isn't this enough? Over a piece of sentence or out of breath, change it. Your ears pick up what your eyes do not see.
- Include a vignette or observation of your own – even a sentence makes all the difference.
- Cut the introductory section. AI is very fond of warming up before making their point. A good proportion of that warm-up can be expended.
Mix up the length of your paragraphs. One sentence can constitute a paragraph. So can eight. It is important to attempt something different from time to time.
- Use - ( ), : - they achieve the punctuation rhythm of a human voice (and actually are useful for emphasis).
- Is this something I would say out loud? If not, pass it through a couple of times until you would.
Knowing your audience and maintaining consistency in your voice - this part is more important than most people realize. AI created content is inherently broad. You are seemingly providing information to everything that is, at the end of the day, trying to cater to no one.
Making content more human isn't about editing (correcting tenses, avoiding clunky language, and other rules): It's got to do with understanding your audience and ensuring whether you are truly conveying this understanding.
A financial blog directed toward newly graduated college students should be totally distinct sounding from a blog aimed for elderly investors. Whenever writing a product description for a skateboard company, it shouldn't read like a memo.
Which they might see as obvious: But when editing output from an AI, it's easy to lose sight of what the text is trying to blend into, rather than reading as an abnormally resonant tone in itself.
Another aspect related to it is voice consistency. If the tone of your brand is conversational and straightforward in your e-mails, but stuffy and formal in your blog posts, it will be perceived. Readers may not be able to say exactly how, but they will feel it.
Create a voice guide to review before editing any AI content: vocabulary you like to use, acceptable contractions, average sentence length, level of humor, etc. Then apply this as a filter to AI output when you are editing.
To be honest many of the humanizing tasks in good AI writing are just plain old editing—as any fine writer has learned. AI provides an initial version. Your task is to help make it sound like you, or at least the image of you that you're trying to project by way of the brand you're representing.
Best correction practices for AI content
When you see an AI response, do a coarser and a finer edit. Check the [AI responses] tab for the actual AI response. Has the AI: a. Attempted to answer the question (or the one that was asked)? b. Answered the question (or the one that was asked) accurately? (e.g., by correcting spelling errors or formatting inconsistencies)?
Please revise your writing throughout, not just in one effort. Three passes: 1. for everyday standard E-Mail structure, tone (how you sound), and clarity: 2. for word choice and details., 3. corrections for spelling and grammar.
There are also other techniques, which are not easy to implement, of speculating about the differences between the data.
Use a (preferably automated) method that will delete 20% or more of the original text. For example, AI is very verbose, or AI will tend to delete less than 20%. Eliminate mercilessly.
Changing the legal entragroup. The legal entragroup will be changed into a wording, referring to provisions within the legal entoswaggroupe. All the people working here will be included and the concept of life has to be frame in terms of the legal entagroup. (Thank you.)
Replace each very, extremely and incredibly with a specific detail or eliminate entirely. By the notion that passive and elaborated if the grand narrative of their own time and more generally of civilization.
Additional points to consider that have not been discussed in the original: Details of the policy were not easily available (providing nobody thought of copying it!). For instance, it was not obvious whether students had options for their oral presentation topics.
Not only the use of visual symbols in Japanese commercials differently influences consumers, they are also differently conceptualized by their users. The traditional Japanese belief that visual symbols have a generic, bottomless meaning, may account for differences in how they are analyzed.
Check the opening sentence of the section. AI generally begins at a very weak state. Nash it on the membrane. Here's an engaging title to draw readers in right from the start.
As a side note, utilizing the Camera feature on your iPhone when playing words can hinder your ability to hit ends; therefore, it's preferable to save this until after you have successfully made calls.
Finish this and leave the audience with something they will remember - a powerful conclusion, an obvious point or a straight out honest opinion. Teens usually end AI with rather weak generic conclusion.
Final thoughts on humanizing AI content
Camouflaging the fact that you used AI to help you write, it's not about hiding it. It is about a valuable first draft and human edits: make it specific, make it personal, give it a tone, make it sound alive, make it sound like you understand who your reader is.
The volume and structure of the AI. You take voice and soul. That kind of division of labor, so long as you do it properly, will yield content that is both efficient and actually good.
Come on... and honestly not to mention, you can't beat that with a stick.
