Pangram in Substack? I’m so excited for the AI witch hunts

So Substack just dropped an unwanted feature on us: “Going forward, you’ll be able to scan text to see how much of it was likely written by hand or with AI assistance,” says Chris Best in Against Claudefishing. Well thanks, Chris. Didn’t realise I needed to be harassed even moar by anti-AI trolls. Trolls, whom, I question their critical thinking skills day by day because you totally hate people using AI but you’re using AI to hate on people using AI? Hilarious. Anyway, if you haven’t gotten the memo, despite assertions that they’re much better now, AI “detection” tools are notoriously unreliable. I had a conversation with my lecturer friend yesterday and I asked her if she ever had false positives/negatives with her students. Without hesitation, she said: “Yes. Definitely!” Fortunately, she belongs to a college where students are not automatically penalised just because an AI tool says they wrote with AI. Her students are interviewed about their essay and asked to defend their thesis.…