You spent countless hours researching, writing, and perfecting your essay, only to have it flagged as potentially AI-generated by your professor’s detection tool. A cold feeling of dread washes over you. How is that possible? Your words are original, the ideas uniquely your own. Or are they?
With the rapid advancement of language models like GPT and Claude, the line between human and AI-generated content is becoming increasingly blurred. These LLMs can craft remarkably coherent essays on nearly any topic imaginable. But can they fool Undetectable AI?
Yes, you heard that right — Undetectable AI isn’t just an AI bypasser, it also has an AI detection tool. In this article, we’ll determine the accuracy of Undetectable AI’s ChatGPT Detector against AI-written text. Stay tuned!
What is Undetectable AI?
Undetectable AI is an AI bypassing platform that features a paraphrase that can hide AI writing from detectors. Apart from having a robust AI rewriter, it also has output customization features to make sure that its output doesn’t steer too far away from its purpose and reading level.
The pricing for Undetectable AI varies depending on whether you choose monthly or annual payments. For monthly plans, costs begin at $14.99 per month for 15,000 words and increase to $209 per month for 380,000 words. Alternatively, annual plans start at $5 per month for 10,000 words, with pricing scaling up to $89.92 per month for 300,000 words. (Note: Prices may or may not have changed after this article was published)
Undetectable AI has always been a favorite of ours and it’s a respect that’s well-earned after we’ve reviewed it extensively in the past. Spoiler alert: it never let us down. Here’s our official review of Undetectable AI.
What is Undetectable AI’s ChatGPT Detector?
We have reviewed, tested, and written about Undetectable AI in the past. In fact, a huge chunk of our catalog highlights how effective Undetectable AI is at avoiding AI detection. But we rarely talk about its AI detector, so that’s what we’re going to do today.
But here’s the kicker: Undetectable AI technically doesn’t have a detector. So, what it does instead is take the average AI likelihood scores from the following AI detection tools:
These detectors are among the best in the market, and we have the article to prove it. However, there’s one that stands out here in my eyes: OpenAI. The thing is, OpenAI actually discontinued their AI detector after they admitted that AI detection is a fool’s errand at this point in time. So, I have to wonder — how does Undetectable AI actually use a detector that’s no longer available?
Unfortunately, I have no answer for that question — still, it’s a question worth asking if we’re talking about transparency and trust.
But that’s not why we’re here today. The only question that matters is…
How Accurate is Undetectable AI’s Detector Really?
Test #1
Undetectable AI: Uncertain about the text’s origin.
AI Likelihood Score: 50%
Test #2
Undetectable AI: Successfully detects text as AI-generated.
AI Likelihood Score: 56%
Test #3
Undetectable AI: Successfully detects text as AI-generated.
AI Likelihood Score: 88%
Test #4
Undetectable AI: Successfully detects text as AI-generated.
AI Likelihood Score: 56%
Test #5
Undetectable AI: Successfully detects text as AI-generated.
AI Likelihood Score: 94%
Test #6
Undetectable AI: Uncertain about the text’s origin.
AI Likelihood Score: 44%
Test #7
Undetectable AI: Successfully detects text as AI-generated.
AI Likelihood Score: 56%
Test #8
Undetectable AI: Successfully detects text as AI-generated.
AI Likelihood Score: 100%
Test #9
Undetectable AI: Successfully detects text as AI-generated.
AI Likelihood Score: 56%
Test #10
Undetectable AI: Successfully detects text as AI-generated.
AI Likelihood Score: 88%
Test #11
Undetectable AI: Successfully detects text as AI-generated.
AI Likelihood Score: 100%
Test #12
Undetectable AI: Successfully detects text as AI-generated.
AI Likelihood Score: 100%
Overall Tally
|
Undetectable AI’s Accuracy Score |
|
A Short Discussion About AI Detection
As LLMs like ChatGPT become more popular, schools and teachers are understandably concerned about students using them to cheat on assignments. As a response, many have turned to AI detection tools that claim to identify AI-generated text.
Sure, Undetectable AI is a great choice for AI detection, but it doesn’t change the fact that current AI detection methods are highly unreliable.
They can be fooled by relatively simple techniques like having the AI introduce intentional errors or paraphrasing its output, which is what Undetectable AI’s paraphraser actually does. Simple editing and adding figures of speech can also make a world of difference to your AI text when it’s processed by a detector.
As AI language models rapidly improve, their outputs will become virtually indistinguishable from human writing. So, what can we do?
Rather than detection methods, teachers should focus on reinforcing good writing practices and redesigning assignments to be AI-proof. Asking students to show their work, cite sources, or complete assignments in controlled environments can mitigate AI use more effectively than any detection software — Undetectable AI included.
All Said And Done
As a fan of Undetectable AI, it’s good to know that it’s not only a capable AI bypasser, but an effective AI detection tool as well.
But as I mentioned, don’t take its word as gospel. No matter how good an AI detector is, there’s no escaping false positive AI detection. The responsible thing to do is to use this as a supplementary tool, but always do your own research to accurately determine if a piece of text came from ChatGPT or not.
If you’re interested in knowing how well it compares against other AI detectors, here’s our latest testing of all the popular AI detectors in the market. Have fun reading!
