AI-generated content is running rampant everywhere, and it seems like there’s a new detection tool popping up every day — each claiming to catch even the sneakiest AI-written text.
But how well do they actually work? After all, let’s face it: AI detectors are a dime a dozen nowadays, and most of them are just smoke and mirrors.
So, we’re taking a closer look at Writer.com and putting it through its paces. Spoiler alert: it’s not as simple as it sounds. Stick around as we dive into the good and the bad (well, mostly bad, but let’s save that for later).
What is Writer.com?
As the name suggests, Writer.com writes.
To be more specific, Writer is an AI writing platform made to change the way businesses create content. At its core, Writer aims to solve a common challenge faced by many organizations: maintaining consistency and quality. With LLMs, Writer allows companies to produce on-brand, high-quality content at scale.
One example of Writer’s use case is that it ensures that all content created through the platform aligns perfectly with the organization’s brand identity using its ability to learn and adapt to a company’s unique voice and style guidelines.
Writer.com is more than automation — it’s about making your life as a business owner easier. They offer a range of AI assistants that can help with tasks like generating ideas and improving readability. By taking care of these more routine aspects of writing, Writer frees up content creators to focus on strategy and creativity.
But we’re not here to talk about Writer as a whole. Instead, we’re going to focus on…
Writer.com’s AI Detector
Apart from their writing features, Writer.com also has an AI detection tool that they offer for free (for 5,000 words). Like most free detectors, they don’t have much in terms of features, but they do offer it as an API.
Can It Detect Content From Popular LLMs?
Like I said, Writer.com’s AI detector doesn’t have any features. This means that the only way we can measure its viability is by determining (to a certain degree) its accuracy. So, let’s test multiple writings from popular LLMs (ChatGPT, Claude, Gemini) against Writer.com.
ChatGPT
Writer.com wrongly classified ChatGPT text as human.
AI Likelihood Score: 26%
Writer.com wrongly classified ChatGPT text as human.
AI Likelihood Score: 30%
Writer.com wrongly classified ChatGPT text as human.
AI Likelihood Score: 25%
Claude
Writer.com wrongly classified Claude text as human.
AI Likelihood Score: 34%
Writer.com wrongly classified Claude text as human.
AI Likelihood Score: 25%
Writer.com wrongly classified Claude text as human.
AI Likelihood Score: 18%
Gemini
Writer.com wrongly classified Gemini text as human.
AI Likelihood Score: 40%
Writer.com wrongly classified Gemini text as human.
AI Likelihood Score: 37%
Writer.com wrongly classified Gemini text as human.
AI Likelihood Score: 43%
Overall Tally
Can It Detect Content From Undetectable AI?
Now that we have Writer.com’s average AI likelihood determinations for LLMs, let’s talk about Undetectable AI. This AI paraphrasing tool acts like a humanizer, wherein it takes an input that’s easily detectable as AI and turns it into something more human.
Think of this as the hardest test an AI detection tool can take. After all, most detectors failed to pass.
If you’re interested in learning more about Undetectable AI, head over to our complete review of the platform or check out any of our comparison articles of it against established paraphrasers in the market.
Test #1
Writer.com wrongly classified Undetectable AI text as human.
AI Likelihood Score: 1%
Test #2
Writer.com wrongly classified Undetectable AI text as human.
AI Likelihood Score: 0%
Test #3
Writer.com wrongly classified Undetectable AI text as human.
AI Likelihood Score: 0%
Test #4
Writer.com wrongly classified Undetectable AI text as human.
AI Likelihood Score: 0%
Test #5
Writer.com wrongly classified Undetectable AI text as human.
AI Likelihood Score: 0%
Overall Tally
What Are The Pros and Cons?
|
So, What Now?
I’ve always believed that I’d rather have AI detection tools that are more lenient than strict. But man, not this lenient. At this point, why even have an AI detection tool if it doesn’t work?
Harsh words — but I stand by them.
Just for reference, we already did an in-depth testing of the most popular AI detection tools a few months ago. No AI detector in that list failed to classify at least 4 AI-generated content except Writer.com, who only managed to successfully identify 2 out of 12. This round of testing was even worse since it didn’t even manage to get a single one right.
This is the point where I usually give positive feedback just to even things out but I don’t really think that they have any redeeming factor. Since Writer.com has been very active in improving their core features, my only hope is that they share some of that love towards their AI detection algorithm as well.
