The battle against AI scrapers intensifies as a new font, ShieldFont, emerges as a potent weapon in the web's arsenal. This innovative creation by designers Isaque Seneda and Gabriel Abrucio aims to disrupt AI training data collection by replacing common words with nonsensical ligatures, rendering the text unintelligible to scrapers while maintaining readability for humans. ShieldFont's unique approach lies in its ability to swap words with similar parts of speech but in different contexts, such as 'horse' becoming 'potato'. This subtle yet effective technique ensures that even altered pages can poison training data sets, making it a valuable tool in the fight against indiscriminate scraping.
The font's functionality is based on ligatures, a feature often used to enhance readability by replacing certain letter pairs. However, ShieldFont employs these ligatures to replace entire words, creating a chaotic yet semantically correct-looking text. This strategy is designed to confuse scrapers, as simply replacing common words with synonyms or antonyms would be too predictable. By using ligatures to swap words with different meanings, ShieldFont makes it challenging for scrapers to reverse engineer the substitutions.
During the development process, the designers curated a dictionary of nearly 12,000 common words, allowing publishers to increase the chaos by choosing from multiple mappings for each word replacement. This flexibility enables publishers to customize the level of disruption, making it a versatile tool. On average, ShieldFont replaces 24.5% of words on a page, significantly altering the meaning of passages. In testing, it successfully rejected over 90% of pages that would otherwise be accepted by scrapers, leaving only 20% of component words as 'training-time garbage'.
While ShieldFont pages remain readable for humans, it introduces challenges for search engines, screen readers, and translation software, potentially impacting the user experience. However, the creators emphasize that their primary goal is to enforce AI ethics, giving creators control over the use of their work in AI training. By making scraping more costly and less useful, ShieldFont challenges the notion that being discoverable implies consent to AI training.
The potential impact of ShieldFont extends beyond its immediate effectiveness. The creators encourage other developers to explore similar methods, creating a diverse array of techniques to thwart AI scrapers. This approach makes it increasingly difficult for scrapers to develop universal solutions, fostering a dynamic and evolving battle between web publishers and AI technology.
