As we are in plain ChatGPT hype, as everyone seems to be aware of this AI tool, one of the questions people ask is, “How to avoid your text being detected as written by ChatGPT by AI detection software“? Reportedly, universities and high schools forbid children from using ChatGPT to produce content. Why would you be willing to write something independently when you can do your essay with ChatGPT?
Many platforms for detecting texts written with ChatGPT or other AI tools exist. These are used primarily by SEO agencies, as their main focus is to avoid their texts being plagiarized or detected as being written by AI tools. If they fail to use these detection platforms, there is a good chance for Google to detect plagiarism or AI-generated content, and the website publishing the content could be penalized, let alone the work in vain.
Open AI had launched an AI detection platform, AI Text Classifier, which also generated hype. Why? Simply put, it is another product from Open AI, which launched ChatGPT. So, who else is better at predicting if ChatGPT writes a text except for the Open AI?
AI Text Classifier was a tool developed by OpenAI that can predict the likelihood of AI-generated text text. It is a fine-tuned GPT model that can analyze text from various sources, including ChatGPT. The tool is likely to spark discussions on AI literacy. However, it will still be primarily used by SEO experts and people who want to detect AI-generated texts, including professors in schools and universities. It was used by AI SEO experts but also by other professionals. We say it was used because it was discontinued recently.
However, it must be noted that AI Text Classifier had certain limitations. For accurate results, it requires a minimum of 1000 characters, approximately 150 to 250 words. Additionally, the classifier was reported as inaccurate, mislabeled AI-generated and human-written text. Additionally, AI-generated text can be easily edited to evade the classifier.
Moreover, the classifier may not perform well on text written by children or in languages other than English, as it was primarily trained on English content written by adults. Therefore, it was essential to be cautious while using the AI Text Classifier and to consider these limitations before making any conclusions based on its results.
Even the company that developed ChatGPT and the AI Text Classifier said the AI detection tool could predict the likelihood of AI-generated text. Still, it was inaccurate and may mislabel AI-generated and human-written text.
What does Open AI say about the limitations of the tool?
We’ve trained a classifier to distinguish between text written by a human and by AIs from various providers. While it is impossible to detect all AI-written text reliably, we believe good classifiers can inform mitigations for false claims that a human wrote AI-generated text: for example, running automated misinformation campaigns, using AI tools for academic dishonesty, and positioning an AI chatbot as a human.
Our classifier is not fully reliable. In our evaluations on a “challenge set” of English texts, our classifier correctly identifies 26% of AI-written text (true positives) as “likely AI-written” while incorrectly labelling human-written text as AI-written 9% of the time (false positives). Our classifier’s reliability typically improves as the length of the input text increases. Compared to our previously released classifier, this new classifier is significantly more reliable on text from more recent AI systems.
We’re making this classifier publicly available for feedback on whether imperfect tools like this one are helpful. Our work on detecting AI-generated text will continue, and we hope to share improved methods.
How can we avoid being detected by AI Classifier or other tools which detect AI-generated content?
If you want to write text using ChatGPT and avoid detection by AI Classifier, here are a few tips to keep in mind:
Write in a conversational style: The classifier was trained on a variety of AI-generated text, which may have a more formal or mechanical writing style. Try to write in a more conversational and human-like manner to evade detection.
Write with software that produces human-like written content: one of the software tools we recommend is WordAI, which passes AI detection.
Mix up your writing style: Mixing up your writing style can make it harder for the classifier to determine if AI generated your text. Try using different sentence structures, vocabulary, and syntax to make your text more diverse.
Avoid using repetitive or formulaic language: The classifier may pick up on repetitive or formulaic language commonly used in AI-generated text. Try to write more naturally and diversely.
Incorporate personal anecdotes or experiences: Incorporating personal anecdotes or experiences can make your text more unique and less likely to be detected as AI-generated.
Write in a different language: The classifier was primarily trained on English content written by adults. Writing in a different language may make it harder for the classifier to detect if AI generated your text accurately.
Remember that the AI Text Classifier is inaccurate and may still mislabel your text, even if you follow these tips. Additionally, AI-generated text can be edited easily to evade the classifier, so there is always a risk that your text may be detected as AI-generated.
How do SEO experts write text with ChatGPT so as not to be detected as written by ChatGPT or AI
Ask for personalized output by offering specific input
The first thing SEO experts do to avoid being detected for using these tools is to ask for personalized texts whenever using the ChatGPT tool. This means you have to ask for specific requests. So, instead of asking, “Write an article on how to write optimized texts with ChatGPT”, you have to offer more input to the tool to generate human-like text.
In this example, you have to ask the tool something like:
Please, write a text in informal language with the title How to write optimized texts with ChatGPT. Write it from a human perspective and offer a readable article, easy to understand in any language.
In a previous article, we showed how to optimize articles using ChatGPT. We showed that Google could easily detect AI-generated text and that you always have to ensure your text is unique and not detectable as written by AI.
At the same time, we showed how you could use ChatGPT for SEO. It’s not a complicated process, but you must thoroughly complete each stage to rank your content.
Offer your structure for the article to be written by ChatGPT
The second thing SEO experts do is structure their articles on topics. This way, you avoid receiving a general text written in a manner easily detected by AI detectors. You have to offer ChatGPT your structure of subtitles, such as for the article to be more personalized.
The output would be a more unique and human-like text that humans can read and understand. Also, you avoid AI tools’ repetitive structures when asked to write on a specific subject.
Come up with your input on the AI-generated text
SEO experts using AI tools never leave articles the way they AI generated them. They develop their expertise on the subject, give examples, and offer their own experience and thoughts. Thus, the text will be more valuable for the readers, and Google will rank your text higher in SERPs.
Besides, the AI-detection tools will find it more challenging to detect AI-written text within the article to be evaluated.
Verify the article with AI-detection software and improve it even more with your own expertise if the case
Always check your article with the AI detection tools, such as AI Text Classifier. If the tool says your text is AI-generated, fix this by offering a more personalized perspective.
Instead of conclusion: can you avoid AI-detection tools?
The answer is Yes, you can. As shown above, there are ways to avoid detection, but, again, any method involves human touch.
There are some methods wholly automated, but they are not free.
Our advice for anyone using AI-generated texts: offer your own expertise on the subject you write about and check your article for consistency, grammar and plagiarism. If the text is valuable, Google will rank it, professors will appreciate it, and you’ll be able to generate even more content this way.
Last but not least, always input your expertise, put down some of your opinions and ideas on the subject and develop a human touch to the content created with ChatGPT.
Proof of success for the method we use? We submitted this article to the AI Text Classifier tool, and the result was that the classifier considers the text doubtful AI-generated.