“PassMe helps you create undetectable AI writing with sufficient human touch.”
We’re back with another AI humanizer review, this time focusing on PassMe.
The quote above is directly from their home page, but is it as effective as it says? Does it actually create “undetectable AI writing?” Let’s find out!
Here, we’ll discuss the following:
PassMe.ai is another AI humanizer that has been heavily promoted on Twitter in recent months, claiming to bypass AI content detection.
PassMe.ai states it can outsmart AI detectors, making content seem more human and helping users quickly generate content.
We’ve seen countless tools make this claim, so we wanted to see the results for ourselves.
In recent times, Google has clamped down on AI usage, penalizing those that rely too heavily on generative AI tools (read more about Google and AI penalties).
For many, that has meant they now spend more time working on their written work to create high-quality content that follows Google’s people-first guidelines.
However, for others, it has meant that AI humanizers have become an intriguing proposition.
Although there is interest in these tools, does this one actually work? Let’s find out!
To see whether or not PassMe.ai actually works, we decided to test it for ourselves.
Step 1: We started the test by generating text within ChatGPT, using the ChatGPT 4 model to get the most up-to-date content output.
Step 2: From there, we copied that text into three AI content detectors; Originality.ai, Writer, and GPTZero.
Step 3: We then took that same copy and pasted it into PassMe.ai, to get our ‘humanized’ version.
Step 4: Lastly, we took that new copy and ran it through the same detectors, to compare the results.
Sample prompt:
"Write an article on organizing a walking tour with friends. Start with the benefits of walking tours, such as fitness and bonding. Provide a step-by-step guide on selecting a route, planning logistics, and ensuring safety. Include tips on making the tour engaging, like incorporating games or themed activities. Conclude with ideas for documenting the experience. Ensure the tone is friendly and the advice is actionable, encouraging readers to plan their own tour."
Original content (ChatGPT-4 generated): 100% Confidence the text is Likely AI
Learn more about AI detection and its efficacy in our AI detector accuracy review.
PassMe.ai Humanizer version: 99% Confidence the text is Likely AI
Try the Originality.ai AI Checker today!
Original content (ChatGPT-4 generated): 100% Probability the text is AI
PassMe.ai Humanizer version: 42% Probability the text is AI
Original content (ChatGPT-4 generated): 17% AI & 83% Human-written
PassMe.ai Humanizer version: 0% AI & 100% Human-written
Despite PassMe.ai’s best attempts, it was unable to trick the Originality.ai detector into thinking it was human-written content.
The Originality.ai AI detector continued to identify the ‘humanized’ text as AI with a high degree of confidence.
Now, let’s take a closer look at what the ‘humanized’ content, actually looks like:
The text has a number of typos and grammatical errors, making it difficult to read and less engaging for an audience looking for high-quality human-written perspectives.
Learn more about AI detection:
Similar to other AI humanizers, you can paste up to 200 words in PassMe on a free plan, before needing to pay to use the tool further.
From our analysis, the PassMe.ai tool isn’t effective at creating human-like content. The Originality.ai AI detector still identifies the content as AI.
How ethically okay it is to use generative AI depends entirely on the circumstances. In most cases, transparency is absolutely crucial to ensure you are not penalized for overreliance on these tools.
Read more about the ethics of AI-generated content.