Anti-AI fonts are useless and harmful

Look, I get the impulse. You pour your soul into a newsletter, a blog, a piece of code documentation, and suddenly every LLM on the planet is chewing through it like a buffet. The idea of a font that makes your text invisible to AI scrapers sounds like a righteous middle finger to the machines. But here's the cold hard truth: it's theater. It's security theater with kerning. The article correctly identifies why these fonts are a waste of time. First, AI detectors are already garbage—they're unreliable, biased against non-native English speakers, and produce false positives constantly. Adding a font that injects zero-width characters or Unicode trickery just makes your text harder for humans to read and easier for anyone with a regex to strip. Second, the accessibility angle is the real kicker. Screen readers, text-to-speech tools, and even basic browser features choke on these obfuscated glyphs. You're not sticking it to the machines; you're sticking it to visually impaired readers and anyone trying to copy your code or quotes. The deeper issue here is that this whole "anti-AI" arms race is a distraction. The open source community has better tools in the arsenal: robots.txt (weak but symbolic), licensing (MIT vs. proprietary matters more than people think), and the quiet power of simply not publishing everything to the public web. If you're really worried about AI training on your work, the answer isn't a clever font—it's deciding what you publish, where you publish it, and under what terms. That's a real strategy. A font that makes your text unreadable to both humans and machines is just a way to make your work useless to everyone. Source: https://blog.yaros.ae/anti-ai-fonts-are-useless-and-harmful/