The web’s newest weapon against AI scrapers is a font
Ars Technica Kyle Orland
A new font called ShieldFont tries to fool AI scrapers with hidden word swaps. Humans see normal text; bots get junk in the HTML.
Based on reporting by Ars Technica, Kyle Orland — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
AI scrapers have pushed publishers into a messy mix of lawsuits and technical countermeasures, and now two designers are taking an oddly elegant swing at the problem: a font.
ShieldFont comes from Isaque Seneda and Gabriel Abrucio, who say in a white paper that it is meant to give web publishers “a practical opt-out from unauthorized AI training” and to “disrupt what is collected when that choice is ignored.” The pitch is simple. Make the page look normal to people. Feed something else to the machines.
The trick leans on ligatures, the old font feature that swaps certain letter pairs for a cleaner-looking version when they sit together. ShieldFont turns that idea sideways. Instead of prettifying text, it uses ligatures to replace whole words with other words, scrambling the value of the page for scrapers.
And the swap only happens when a browser renders the page on screen. If a scraper just pulls the plaintext source, it gets the altered version instead — the one no reader ever sees. That makes ShieldFont less like a lock and more like a booby-trapped signpost: the page still works for people, but the raw material underneath gets poisoned.
My take — AI-written commentary, not fact-checked reporting
This is the right kind of nasty. If AI firms want the public web as free training fuel, the public web is going to start feeding them broken text, fake text, and a few more tricks besides. The real story here isn’t a font; it’s that publishers are finally treating scraping like a bug to be patched, not a law of nature.
Read more about this at: Ars Technica
Related stories
Open source project fools AI scrapers with poisoned font
The Register · 1 month ago ·
46
Anthropic plans to add an invisible mark to AI text—as the industry scrambles to police AI slop
Fortune ·
41
Anthropic’s Text Watermarking Proves AI Companies Do Not Care at All About Writing
404 Media · 4 weeks ago ·
15