We checked if ShadowGPT, a tool meant to make AI seem human, can actually get past smart AI checkers without getting caught.
We tried it with stuff made by ChatGPT-4. Here’s what we found out.
“The only platform that truly bypasses AI detection algorithms. We are not shy about showing real AI detector results like GPTZero.”
ShadowGPT’s main page makes a big and brave claim!
ShadowGPT acts like those AI ‘humanizers‘ that say they can make AI-written content look like it wasn’t written by AI, and fast.
Like we’ve done with other AI ‘humanizers’, we tested ShadowGPT to check if it can truly avoid detection tools.
This Article Talks About :
– A close look at the ShadowGPT that makes writing more human-like.
– The way ShadowGPT does its job.
– Why people are curious about tools like ShadowGPT.
– Trying the tool out on blog writings.
– A final thought on if ShadowGPT’s writing can trick AI detectors.
ShadowGPT is a tool not many people know about. But lately, it’s been popping up on social media. It changes AI-written text so that it won’t get caught by AI-checkers like Netus.ai.
How does it work? You give it something written by an AI, and it changes it.
This makes the text seem like it wasn’t written by an AI, helping it avoid being spotted by AI detectors.
People like using text makers such as ShadowGPT, mainly to create stuff to read.
Google doesn’t want content made by AI anymore. So, many people are going back to writing great content themselves.
Yet, some still make AI clickbait.
Certainly! Below are expanded versions of the two subheaders with additional detail to total approximately 700 words.
AI “humanizers” like ShadowGPT make AI-created words sound more human and real. These tools do a lot of things that change the way AI writing sounds. They can pick different words, form sentences in new ways, and even fix things that don’t make sense. This makes the writing easier to read and understand. But, they are not perfect and sometimes can cause problems.
Good Things About Using AI Humanizers:
Negatives of Applying AI Humanizing Tools:
To really know if ShadowGPT can do its job, we need to test it out. Is ShadowGPT able to make content that seems like a person wrote it, so it doesn’t get caught by AI checkers?
The Test Explained
First up, we made an article with the help of AI – that’s our starting point for the experiment. We used ChatGPT to create this piece.
Next, we took this AI-made article and put it into three well-known AI detection tools. These tools were GPTZero, Writer AI, and a tool we made called Netus.ai AI Checker.
Thirdly, we used the ShadowGPT tool to ‘humanize’ the article. We wanted to see if after ShadowGPT changed it, the AI detection tools could still tell it was written by an AI.
Lastly, we took the newly made ShadowGPT text and checked it again using the same tools to spot any changes from the original text.
Here is our prompt :
“Write an article about selecting the best home office printers for remote workers. Start by discussing the importance of a reliable printer for home offices and outline the key needs it should fulfil, such as multifunctionality, quality, and connectivity. Compare different types of printers (inkjet, laser, all-in-one) and their suitability for various workloads, highlighting advantages and disadvantages. Detail the features to consider, such as print speed, resolution, and cost-efficiency, including both initial and ongoing expenses. Conclude with recommendations for top printer models at different price points, catering to both budget and premium preferences, to help readers make an informed decision for their remote work setups.”
Netus.ai Score
This text was made by ChatGPT-4 and is probably made by an AI with 100% confidence.
ShadowGPT made this text, pretty sure a computer did it with 98%.
GPTZero Score
Original content (ChatGPT-4 generated): 88% AI
ShadowGPT Humanizer version: 74% AI
Writer Score
Original content (ChatGPT-4 generated): 15% AI & 85% Human made
ShadowGPT Humanizer version: 8% AI & 92% Human-written
The results show ShadowGPT can sometimes trick detectors, but it can’t fool the best ones like Netus.ai.
If you take a closer look at the text, you’ll understand why it could trick some AI detectors. Here are a few reasons:
These examples are full of sentences that are hard to understand, spelling errors, and letters that seem out of place. This means that although they might trick some detectors, these texts don’t look like something a person would write well.
Is it free to use the ShadowGPT tool to make text sound more human?
No, the ShadowGPT tool isn’t free. It offers three different prices to choose from. You can decide if you want to pay every year or every month.
Does the ShadowGPT helper work well?
From what we’ve found, the ShadowGPT helper doesn’t really work well.
Is it okay to use AI to help with writing?
Talking about if using AI for writing is okay is something many people do. Most of the time, people expect writing to be done by humans, but sometimes AI is used to help.What should you do? It’s important to talk about the rules for using AI and agree on them before starting to write. This way, everyone knows what is expected.
Through digital leadership we empower people to leverage the opportunities in global economy
@ 2024 Netus AI.