In this blog post, we’ll explore the controversy surrounding AI-generated content, specifically focusing on GPT chat detectors. These detectors are tools that analyze text and determine if it was written by a GPT language model, such as GPT-2 or GPT-3. With the growing use of AI in content marketing and blogging, many are wondering how Google will detect AI-generated content and if it will affect their search rankings. In this post, we’ll discuss how well these detectors work, test them out for ourselves, and explore the claims that originality.ai is a scam.
To test out the detectors, we’ll be using three popular AI detectors – originality.ai, Detect GPT, and the GPT-2 Output Detector. While the GPT-2 Output Detector is a free tool that only detects GPT-2 output, it still works well for detecting GPT-3 output.
I started by testing out the GPT-2 Output Detector with content from the Future tools website. Surprisingly, the description for Detect GPT was written by AI, so we posted it into the detector and received a result of 96.67% real and 3.33% fake. I then tested the detector with content from originality.ai and received a result of 97.56% real and 2.44% fake. However, when i tested it with a prompt generated by GPT-3 in Chat GPT, the detector was unable to detect that it was fake.
Next, I tested Detect GPT with the same originality.ai content and received a result of 97.56% human and 2.44% AI. However, when i tested it with Auto draw’s AI-generated description, it showed a result of 57.78% AI and 42.22% human, indicating that i’m 57.78% confident that the content was generated by AI.
Finally, I tested the detectors with 100% human-generated content from AI business blog, and all the detectors were able to detect that the content was real.
In conclusion, the detectors i tested were not completely reliable when it came to detecting GPT-3 generated content. However, they worked well for detecting GPT-2 generated content and were able to detect 100% human-generated content accurately. As for the claims that originality.ai is a scam, i’m unable to find any evidence to support this, and it’s possible that the Reddit thread was removed for other reasons. Overall, it’s clear that the use of AI in content marketing and blogging is on the rise, and while these detectors may not be perfect, they can still be a helpful tool for detecting AI-generated content.