DEVOURED
This Font Looks Perfectly Normal to Humans but Wreaks Havoc on AI
ShieldFont manipulates HTML ligatures to display decoy text to AI scrapers while rendering the intended message correctly for human readers.
What: ShieldFont uses typographic ligatures to swap characters in the DOM, forcing LLM scrapers to ingest incorrect data while the browser renders the intended content visually for users.
Why it matters: This represents an escalating arms race in web scraping, where developers move beyond standard robots.txt restrictions toward active content-spoofing to prevent AI training on proprietary data.
Deep dive
- Ligature substitution: Uses specific character combinations that render as different words in the browser but appear as raw, decoy text in the source code.
- AI resilience: Targets the tokenization process of LLMs by injecting noise directly into the training data stream.
- DOM integrity: Operates at the font layer rather than the CSS or JS layer, making it harder for simple headless browsers to detect via basic inspection.
- Visual parity: The technique is designed to be invisible to end-users who see clean, readable text.
- Scraper mitigation: Offers a novel way to protect text content from automated ingestion.
Decoder
- Ligature: Two or more letters that are joined together to form a single glyph.
- Scraper: An automated program that extracts data from websites.
Original article
ShieldFont uses ligature-style substitutions to swap whole words in a page's HTML source, so scrapers collect decoy text while human readers see the original.
Explore all 58 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi