Member-only story
How to Fool AI Detectors Easily
AI-written content is becoming increasingly common, and the need to ensure it passes detection systems is growing. Whether you are writing a blog post, an essay, or an article, making your work feel human is critical. That is what’s referred to as humanised content in this season where AI-written content floods everywhere. If you have ever wondered how to trick AI detectors, a tool called Bypass GPT claims to help your AI-generated content slide right under the radar—even when checked with Turnitin and Originality.ai.
Non Medium Members can read this article here
To put this to the test, I created a blog post with GPT-4. I prompted the GPT to make it fun, friendly, and slightly humorous. The idea was to make it sound like something you would read on a casual blog, you know. The question on my mind was, would the generated result bypass AI detection tests?
So I tried it out, and here is what was observed:
I ran the original text through some AI detectors, including GPTZero and CopyLeaks. It was flagged as 50% AI-generated. Some tools passed it, but others, like Turnitin and Originality.ai, needed to be more convinced as they could detect it’s AI-written. So I needed to make further changes to the article.
I selected the Enhanced Mode in Bypass GPT after pasting the AI-written article. I…
