![\"A](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/07\/1784724138_355_0x0.jpg\")
<\/p>\n<\/p>\n
In tests, leading AI models struggled to decode the moving message in Ghost Font’s field of dots until prompted exactly what to look for.<\/p>\n
Eric Lu<\/p>\n
Artificial intelligence, it seems, is always hanging around, ready to interpret and summarize what we write, even if we don\u2019t ask it to. An experimental visual communication system called Ghost Font aims to stand in AI\u2019s way. Computer scientist Eric Lu devised it to make it hard for AI models to decipher messages without human guidance. <\/p>\n
While Lu calls Ghost Font<\/a> an \u201canti-AI font,\u201d it isn\u2019t a traditional font that uses solid lines to form letters. Instead, it represents text using thousands of tiny moving dots that drift in the opposite direction from those in the surrounding background. Because humans can readily resolve patterns in motion \u2014 when points move in the same direction, the brain groups them into a unified whole \u2014 the moving dots appear as faint letters. <\/p>\n When Lu tested leading AI models Claude Fable 5<\/a> and GPT 5.6 Sol<\/a>, however, they struggled to decode the writing. <\/p>\n \u201cI was inspired to make Ghost Font one day after texting with my mom, and I noticed that AI kept trying to summarize our messages, many times incorrectly,\u201d Lu, a former Google associate product manager, said in an interview. \u201cI was annoyed that it was being so intrusive in our private conversation and I started brainstorming ways that I could potentially communicate without AI reading everything.\u201d <\/p>\n For now at least, he sees Ghost Font as less a practical tool than an experiment that highlights an emerging tension as AI assistants increasingly summarize emails, texts and documents, often by default<\/a>. Some users are happy to accept AI\u2019s help, but others want more control over when algorithms can \u201cread\u201d and sum up what they write.<\/p>\n Lu believes Ghost Font could help measure the ever-shifting capabilities of AI, which is advancing at a head-spinning rate. <\/p>\n While in July 2026, it might be difficult for the most capable AI systems available to read Ghost Font without human intervention, \u201cit will be interesting to compare in a few months or next year,\u201d Lu said. \u201cThere could be more video native models that read Ghost Font right away just like a human can.\u201d <\/p>\n One of Lu\u2019s big takeaways: \u201cThe gap is closing, quickly.\u201d <\/p>\n Ghost Font Goes Viral <\/p>\n Lu\u2019s X post about the experiment has so far racked up 17.7 million views, reflecting fascination with the idea that humans might briefly have an advantage over AI and drawing people eager to test its limits. A number have set out to prove that Ghost Font is no match for AI by sharing the prompts they used to get AI to read the typeface successfully. <\/p>\n \u201cI got GPT-5.6 Sol to read \u2018Ghost Font’… by just telling it in what direction it looks like the noise is moving,\u201d Riley Goodside, a former Google DeepMind prompt engineer, wrote on X. <\/p>\n But Lu stresses that he designed Ghost Font to function without being instructed specifically what to look for, and that those trying to prove an AI agent can read it \u201cmiss the point of what the project was about in the first place.\u201d <\/p>\n \u201cThe goal was never to create a completely secure messaging technique that is impossible for any AI to crack,\u201d he said. \u201cThere are plenty of more secure ways to protect messages from AI \u2014 for example, placing a message behind a knowledge factor like a password.” <\/p>\n As long as the data is publicly available, technically, it can be read with the correct prompting, he said. \u201cBut slowing down AI agents can still be useful in many situations,\u201d he added. <\/p>\n A Bifurcated AI Discussion<\/p>\n Lu founded Mixfont, an AI font generator that creates typefaces in seconds from a single prompt or image. He believes discussions around AI tend to be divided into \u201cAI is good\u201d or \u201cAI is bad\u201d camps, and would like to see more nuanced conversations. <\/p>\n \u201cThe project is still valuable as an example of how humans perceive things differently from AI,\u201d he said. \u201cThere are still things, like instant motion perception, that make humans unique, and those avenues are worth exploring and researching.\u201d<\/p>\n ForbesWe Trust AI Over Our Own Brains, Research FindsBy Leslie Katz<\/a><\/p>\n He also believes the project underscores AI fatigue<\/a>, the pressure and expectation workers feel to keep up with the latest AI tools. <\/p>\n \u201cEven as the technical gap closes,\u201d he said, \u201cthere\u2019s still a pent-up desire for conversations and connections that are human-to-human and don\u2019t involve AI.\u201d <\/p>\n