How To Avoid Getting a False Positive AI Detection

On the 30^th of November 2022, OpenAI revolutionized the word of writing, blogging and content creation with their release of a free version of Chat-GPT, other key players like Alphabet, Meta, and Microsoft also contributed to what resulted in the proliferation of generative Ai content all over the internet. With a simple prompt, an Artificially intelligent Algorithm can craft out an essay, a poem or story with impeccable accuracy and speed.

While this had a lot of positive applications, it also created a lot of tension in areas where the human intellect was expected to be used in the creation of content.
As a safeguard to prevent writers form submitting AI generated content as their work, the introduction of AI Detection app and sites became necessary.

What are AI Detectors

AI Detectors are software tools and other counter measures designed to scan a text with the aim of determining whether it was generated by an AI or a human.
Some outstanding examples of AI Detectors you can use to identify content created in whole or in part by an Artificially intelligent program are listed below.

Originality AI	AI Text Classifier
OpenAI Text Classifier	GPTZero
GLTR	Copyleaks
Sapling AI Detector	CrossPlag AI
Writer AI	Content at Scale

These programs use various approach and techniques to come up with a percentage probability that the text was created by an AI. But can these ever be wrong?

False Positive Results

A false positive is a situation where an AI detector concludes that a writeup in it's entirety or partially was composed by an AI when it was actually written by a human.
This is a situation most of us already anticipated. I could see it coming, and feared that I could be cought in "friendly fire" so I avoided content that would display or require plenty of intellectual prowess... Yet, I have seen a number of simple entries flagged as AI generated by a number of AI Detectors.

Why would an AI detection tools wrongfully flag your entry as Ai-created content?

AI Assisted Editing
If you make heavy use of AI tools to edit your content after creating it yourself, using tools like Grammarly or Hemingway to proofread and edit grammatical errors, editing, and adjust sentence structur, you end up having the signature of an AI in your work which could be detected when scanned.

Vary your content’s structure.
Avoid a writing structure that is predictable as that is often the case with AI generated text, they often have paragraph styles that are similar in lenght.

Let your unique signature show
Every writer has a unique writing style like a finger print, it may be the use of certain words or group of phrases that would uniquely single out your work. It is much like a handwriting.

Interpretation with AI
There is a high risk of getting flagged wrongly if you composed your writeup in another language and later had to use an AI to translate it into English. During translation, the translating AI may give you an output which when scanned has the characteristics of an AI generated content.

Test with an AI detector
After writing, subject the work to be scanned by an AI detector application, just to be sure. If you get a false positive, work again on the test and rescan it. Avoid depending on text Suggestions when typing
Here are some example of AI detector you may use.

CrossPlag.com

Copyleaks.com

openai-detector

Zero GPT

Conclusion

Avoid getting too much help from AI when proofreading your content, allow your writing signature to show on your work.
Use the scanner on the content before posting it. Try about two or three of them.