The web’s newest weapon against AI scrapers is a font
Summary
Two designers created a new font called ShieldFont to protect websites from artificial intelligence (AI) scrapers that copy web content for training data. The font shows normal text to human readers, but behind the scenes it changes words to nonsense in the webpage code, which confuses AI scrapers and lowers the quality of their data.Key Facts
- ShieldFont uses special letter combinations called ligatures to swap whole words in the webpage’s code without changing what appears to users.
- The font replaces words with others that are similar in part of speech but have unrelated meaning, for example swapping “horse” with “potato.”
- About 24.5% of all words on a page are replaced, affecting nearly half of the important content words.
- This substitution creates false or confusing information in the data that AI scrapers collect, which can lower the quality of AI training.
- ShieldFont offers three different swapping options per word and allows publishers to customize changes paragraph by paragraph.
- Tests show that over 90% of scraped pages processed with ShieldFont are rejected by AI scraper filters due to poor quality.
- Even the pages that pass the filters contain a significant amount of “training-time garbage,” meaning nonsensical or incorrect English.
- The font helps web publishers opt out of unauthorized AI training by disrupting the data collected by scrapers.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.
