In an attempt to reverse the roles between humans and machines, the font design company Mixfont has created an experimental font that makes the human reader see one message while artificial intelligence systems might read a different, misleading message.
The innovation, according to the Smithsonian website, is called “Decoy Font” and relies on an optical illusion that combines two letters in the same space, one of which appears with thin, clear lines, while the other is formed from a dark, low-detailed mass. The human brain can distinguish between the two layers, especially when the viewing distance is changed, but some image reading models focus on the clearer layer and fail to extract the hidden message.
The font is presented as an experiment to protect human writing from automated compilation programs and text extraction tools, and it can be downloaded in “TTF” format and used within design programs.
However, this method does not provide guaranteed protection and is not a substitute for encryption. The system that accesses the original digital text can read it directly, and newer models may be able to bypass the trick if they know in advance how it works Independent tests indicate that giving artificial intelligence a simple hint may help it uncover the true message.
Therefore, the idea seems more like an experiment in the difference of perception between humans and machines, and a temporary race between the design of illusions and the development of computer vision, than a permanent digital lock.