0 comments
If you’re a student who’s recently submitted work containing traces of AI, you’re probably cooked.
It all seemed a little too good to be true, because it was.
If you were a student who graduated in 2024, you were in the absolute sweet spot for generative AI.
The explosion of chatbot generators like ChatGPT and Gemini didn’t initially coincide with plug-ins to detect where the tech had been used, and for a moment, it felt like academia had no answers. Assignments could be knocked out in 10 minutes and markers were none the wiser.
That window has now slammed shut, however.
In recent months, a wave of tools like GPTZero have sprung to life identifying when and how text works were assembled, including how many times content was pasted into a document, and even which model probably did the heavy lifting.
View this post on Instagram
It’s less ‘this might be AI’ and more ‘you copied this at 2:08am using GPT 3.5 from your iPhone’. If you’ve used an online document, timestamps are even able to show how text arrived in real time, meaning huge blocks over typed sentences won’t go unnoticed.
It’s a funny dichotomy. ChatGPT will generate your work with the nuance needed to secure specific marks, while GTPZero endeavours to narc you out with extreme precision – and without the mercy of a weak human.
It’s not just GPTZero, either. Plenty of detection apps are being baked directly into Google Docs, letting markers scan work without ever opening another tab. Originality.ai throws in a plagiarism score just to twist the knife, and Copyleaks boasts a 99% success rate at flagging suspect sentences in real time.
There are extensions for Chrome, integrations for Google Workspace, buttons inside learning portals. If you’ve written an essay recently, someone probably ran it through something.
