We’ve started talking a lot about AI content detectors at InfinitePeer since there are so many GPT-based AI content providers popping up every week.
Not every piece of content produced by AI is guaranteed to be accurate. AI content is not always factually correct and has the potential for error.
Because of this, anything produced by AI must be reviewed and revised by humans before it’s made public.
How can you determine if a human author or AI software is responsible for a piece of writing?
Making use of software designed to evaluate and identify this content is one technique to determine if a text was produced by artificial intelligence. Here, we review Content at Scale and evaluate if it truly can detect if content is human- or AI-generated.
What is Content at Scale?
Unlike Originality.AI and other AI detection tools, Content at Scale did not start as a plagiarism checker. Content at Scale sells content that they claim passes AI content detection. They claim they are not an AI writing assistant but “a human-level long-form blog post-producing machine!” (What a mouthful!) It seems they really can grind out the words!
Content at Scale doesn’t provide much information about its content detection software, so we’ll need to test this out for ourselves and see just how accurate it is.
Content at Scale does have a decent amount of information on their site about their “Done-for-you services” where you provide them with keywords and they provide you with high-quality blog posts.
Key Features of Content at Scale
Again, Content at Scale’s features is focused more on their paid services, as opposed to their free AI content detector. Focusing on their AI detection service, these are the key features:
- API integration: Upon request, API integration seems to be available. You’ll need to contact Content at Scale to describe your use case though.
- One language: It works in English only
- Free: Totally free.
Signing up for Content at Scale
To use Content at Scale’s AI content detector, you don’t need to sign up. You can access the tool directly from their website without an account.
It’s very simple to use. Just copy and paste your content into the text box and click “Check for AI Content”.
How accurate is Content at Scale AI detection?
If you read our other AI content detection reviews, we tested each tool by putting them through six tests.
- We generated 3 pieces of AI-drafted content using ChatGPT.
- We (us humans!) then sourced 3 pieces of content, two from our website and one from a Forbes article.
We standardized the content we used to truly compare the quality of each product against each other. Ideally, the ChatGPT content would be identified as 100% AI, while the human-written text would show as being 0% generated by artificial intelligence.
Let’s see whether Content at Scale passed the test.
AI-Written Sample One
Success! Content at Scale tells gives this piece a 0% human content score and correctly identified this as “obviously AI”.
AI-Written Sample Two
Somewhat pass! Content at Scale rates this as 76% Fake and 24% Real. In reality, this is 100% AI.
AI-Written Sample Three
Success! Again, Content at Scale identified that this is 100% AI-written.
Human-Written Sample 1
Next, we tested the human-written samples. Content at Scale tells us the first one “looks great!” and even includes a stars-eyes emoji. Success!
Human-Written Sample 2
Sample number 2 is also a success! Although this time around Content at Scale only thinks 99% of my own work is human-written, I wonder what part it thinks I didn’t write?
Human-Written Sample 3
For our third test, we used an excerpt from this Forbes article. Another success! Content at Scale tells us this is 100% human-written.
Overall, Content at Scale was effective at identifying AI vs. human content. For 5 out of 6 samples it correctly identified the author.
One item to note on Content at Scale is that this tool does not also detect plagiarism (unlike tools like Originality and Copyleaks). So even though all human-written samples were correctly identified as human-generated, they were all plagiarized, and this was not flagged. To be fair, the tool doesn’t claim to detect plagiarism, but it’s worth noting that other tools can identify both.
Can Content at Scale be fooled?
After playing around with several AI content detectors, we’ve found some ways to “fool” these tools. Some ways we’ve successfully passed AI content off as human content include:
- Makine spelling mistakes
- Paraphrasing content or changing sentence structures
- Changing punctuation, or including incorrect punctuation
So, let’s see if we can fool Content at Scale. We took the first sample, which Content at Scale previously found to be 100% fake, and made a few tweaks. We included two spelling errors, a few changes in punctuation, and paraphrased some sentences. Then we ran the updated text through Content at Scale.
Uh oh! Content at Scale was fooled, telling us this piece was “looking good!” and even including a heart emoji. They analyzed this content as being 87% real and 13% fake. Given the minor changes we made, if anything, this content is 1-2% human-written.
The Content at Scale tool is completely free. There may be charges for APIs, but as an individual user, there is no charge for the tool.
Content at Scale shows that it can pretty reliably detect AI-generated content. Given the basic features (i.e., English only, no ability to save, up to 2,500 characters only, no plagiarism detection), we recommend Content at Scale for individual users who need to do quick checks for AI content.
For businesses, writers, or professional content creators, we recommend a more robust tool that allows for plagiarism detection as well as other features.
If you have experience with Content at Scale, please write to us! We’d love to hear about our experiences and incorporate the feedback into this Content at Scale review.