I am once again begging people to realize that AI checker doesn’t work. it’s never worked. it’s notoriously known to have flagged human-made works as AI and AI-generated works as human-made. and by feeding it people’s works, you are feeding more works to AI, because apparently the machine itself is AI.
the only thing AI checker does is harm genuine artists and people in general too.
I am once again begging people to critically analyze stuff even when that stuff supports your opinions.
It's easy to laugh about an AI not knowing Frankenstein, but let's look at the rest of the "AI checker" photo, okay? I don't think it's checking AI in the sense of punishing someone for using it. The button at the end reads, "Humanize text." The fine print below the button reads, "Make your text human with undetectable AI" (I edited out the title capitalization, sue me).
This "100% AI GPT" finding is practically an advertisement. The website was made by people trying to sell you something. Specifically, they are trying to sell you a special, magical tool that will let you use AI but not get caught. The target audience for this magical tool consists of people who are already using AI to write shit but who do not want to get caught; rather than just writing the stuff themselves, they are turning to a second AI to "humanize" the AI-generated text so that they can use AI without getting caught. By artificially inflating the % to scary levels - "Oh no," the AI user might think, "my teacher will be able to tell I used AI on this essay if it scores a 100%!" - the seller of the magical tool creates a strong need for their magical tool. "Yes," the seller says, "It's very scary that it scored a 100, but it's okay, I can help you, just pay me money and I'll get that 100 down below <idk, some threshold like 20? 10?>."
The target audience for this magical tool does not include anyone the Isi Breen tweet assumes. I seriously doubt that teachers are using "humanize AI" products for their own students to help the kids not get caught next time, because teachers don't normally teach kids how to get better at cheating. I seriously doubt that employers are using "humanize AI" products to, what, help prospective employees improve their materials for their next job application??
There are absolutely massive flaws in "AI checkers" for academic integrity purposes, and I am sure similar flaws exist in HR spaces. I totally agree that AI and AI checkers are harmful and ineffective, but the above ~Free AI Text Humanizer~ flagging Frankenstein isn't great evidence.























