@theslowai: I ran the opening of Aristotle's Rhetoric through MyDetector, which advertises itself as the most accurate AI detector for content generated by ChatGPT, Gemini and Claude. On 688 words it returned 0% AI, 81% human, and 19% mixed, meaning it reckons roughly a fifth of the text involved AI. Rhetoric was written around 350 BCE, and the translation it read is W. Rhys Roberts's, published in 1924, so even on the most conservative dating these words predate any machine capable of writing by about a century. Two things I did not get to on camera. The passage it flagged is Aristotle arguing that judges should be allowed to decide as few things as possible, because feelings of friendship or hatred or self-interest make them lose any clear vision of the truth: a machine passing biased judgement flagged a warning about biased judgement. And the same site sells an AI Humanizer, with a Humanize the Text button sitting directly beneath your result. One company, both sides of the trade. Text from the MIT Classics archive, and you can repeat the whole thing in about two minutes. Source: https://classics.mit.edu/Aristotle/rhetoric.1.i.html #AIdetection #Aristotle #AcademicIntegrity #EduTok #CriticalAI

Slow AI

Open In TikTok:

Region: GB

Saturday 08 August 2026 18:05:35 GMT