{"id":704759,"date":"2026-07-22T06:31:33","date_gmt":"2026-07-22T06:31:33","guid":{"rendered":"https:\/\/www.newsbeep.com\/uk\/704759\/"},"modified":"2026-07-22T06:31:33","modified_gmt":"2026-07-22T06:31:33","slug":"ghost-font-hides-messages-in-motion-and-ai-models-cant-read-them","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/uk\/704759\/","title":{"rendered":"Ghost Font Hides Messages in Motion – And AI Models Can’t Read Them"},"content":{"rendered":"

Stare at a screen full of tiny dots. Same color, same size, same apparent randomness. Then the video plays. Within two seconds, a word materializes from the chaos, clear as a billboard. Pause the clip. The word vanishes. That\u2019s Ghost Font<\/a>, an experimental typeface built by designer Eric Lu at San Francisco-based Mixfont. It encodes text purely in motion \u2014 and multiple leading AI models<\/a>, including versions of ChatGPT and Claude tested by Lu, have repeatedly failed to read it.<\/p>\n

The project isn\u2019t just a party trick. It exposes a genuine blind spot in how AI processes visual information.<\/p>\n

Motion Is the Message<\/p>\n

Ghost Font doesn\u2019t install on your computer \u2014 it generates short looping videos where letters exist only as movement.<\/p>\n

Here\u2019s the mechanism. Thousands of dots fill each frame. The dots forming a letter move in one direction \u2014 say, upward. Background dots move the opposite way. Because every dot shares the same color and size, a single frozen frame looks like pure static. No edges. No curves. Nothing for OCR to grab.<\/p>\n

Your brain, though, does something AI still struggles with. It applies what psychologists call the \u201ccommon fate\u201d principle \u2014 automatically grouping things that move together. Like picking out a friend in a crowd because they\u2019re the only one walking toward you. The result: you read the word almost instantly.<\/p>\n

\u201cGhost Font uses motion, noise, and decoys to hide messages in videos readable by humans but nearly impossible for AI models like ChatGPT or Claude to decipher.\u201d \u2014 Eric Lu, Mixfont founder<\/a><\/p>\n

I created a font called Ghost Font that only humans can read. Tested it in Fable and GPT 5.6 Sol Ultra and neither was able to decipher it correctly. pic.twitter.com\/wy08KgZKH2<\/a><\/p>\n

\u2014 Eric Lu (@ericlu) July 11, 2026<\/a> <\/p>\n

What makes it tick:<\/p>\n

Letters exist only as motion patterns \u2014 no static shapes anywhere in the videoA paused frame is pure noise; no edges or curves for OCR to grabDecoy text deliberately misleads AI toward false readingsHuman persistence-of-vision fuses frames into letters; AI models typically analyze frame-by-frame \u2014 and Mixfont\u2019s free web playground lets anyone test this gap directly<\/p>\n

Trying to read Ghost Font by analyzing individual frames is like understanding a flipbook by staring at page seven. Models fed these clips either detected nothing, confidently reported the decoy text, or hallucinated words that were never encoded. These kinds of failures are among the most frustrating computer problems<\/a> users and researchers encounter today.<\/p>\n

Clever Illusion, Not a Silver Bullet<\/p>\n

Lu himself acknowledges that a system built to track dot trajectories across frames could crack it.<\/p>\n

Ghost Font is not AI-proof. Any system computing optical flow across frames could, in principle, reconstruct the letters. The text is also harder for humans to read<\/a> than standard typography \u2014 you might need a beat to lock in. This is a current-generation obstacle and a research tool, not a permanent privacy solution<\/a>. As temporal modeling in AI improves, the gap Ghost Font exploits will shrink.<\/p>\n

Still, it went viral for a reason<\/a> beyond novelty. Ghost Font became something like protest art for the scraping era \u2014 proof that human perception still holds territory AI hasn\u2019t fully mapped. That territory won\u2019t last forever, but right now it\u2019s real, and the race to close it is already underway. The broader implications for digital privacy echo concerns raised by every new surveillance app<\/a> built to track people without their knowledge.<\/p>\n