Most AI systems work by reading the pixels of an image up close
In the case of text aren’t LLM’s reading actual text ?Whats the usecase here ? For any text I dont want to be readable by ai scrapers and such, I need to use this font and render as image(s)?
Top points for effort, but if it’s making it hard to read even 2 words… I’d rather poison/prompt inject via invisible text and not make any humans cross-eyed trying to read that font
It works at full resolution, but it’s a cat and mouse game. If something like this becomes popular, then all that needs to be done is to start passing images through a filter first (blur, lower resolution, etc.), or just train a new vision model that can pick out both the blurred message as well as the decoy message.
The downloadable font is especially vulnerable since each letter is a straight map, so easily trainable
Well look at that, it works!!
I’ve found circle to search on android can read both variations depending on how you select the image
Nice ! I didn’t think the font had a serious purpose before I saw your excellent example…
Here’s the text my phone extracted from that image:
FUCK THE POUXCE
What does this say?
It says “THIN BLUE LINE.”
The word BLUE is split between the second and third lines:
THIN
BLU
ELINE
•••
THIN
BLU
ELINEB*
Edge detection moment
Well, I’m glad that we now have a font that makes it look like I’m not wearing my glasses… when I’m wearing my glasses.
Person: does something impressive against AI Comments: it’s bad because it’s not perfect
Good luck to the visually impaired and their screen readers.
Screen readers read text directly, the point of this is to be used in images, which screen readers don’t read anyway.
Facedeer is a bad faith troll and AI evangelist, in case anybody is wondering why he’s pretending to care about accessibility.
Ahh, that explains the willful ignorance.
My iPhone reads text from images. I wonder how it’d do with this?
You can take a screenshot or save the image directly. My iPhone reads the blurry text without issue, so this may be more related to LLMs in particular, or of the image is compressed before the analysis, the illusion is broken
Amazon Rekgnition struggles a little more but it’s still reading the blurry text instead of the lines
And if this sort of font became common (and actually kept working) then screen readers never would be able to read these. The point remains that accessibility is being sacrificed for the sake of some kind of general “anti-AI” movement.
I’m pretty sure they’re not the intended users
This is art
Clever as fuck
I think this is the first step in an arms race between fonts that can’t be read by computers and systems to differentiate between obfuscated text.
Doesn’t work on a dark background ? The decoy text is invisible…
