The New York Times, The Guardian, and USA Today have recently restricted the Internet Archive’s ability to preserve their articles, citing concerns that the Wayback Machine serves as a back door for AI companies to scrape content. This move follows a broader trend identified by Harvard’s Nieman Journalism Lab, which noted that more than 340 local-news sites had limited Internet Archive access as of May 2026. While publishers argue this protects their intellectual property from unauthorized AI training, the action threatens to dismantle a foundational tool for digital preservation and historical accountability.

The Wrong Target for AI Resistance

Internet Archive founder Brewster Kahle argues that blocking the Wayback Machine is a misdirected response to the rise of large language models. Commercial AI companies have already secured training data through direct web crawls, licensed datasets, and proprietary retrieval systems. Kahle notes that the Internet Archive has implemented its own technical controls, including rate limits and filtering, to prevent abusive scraping by AI startups. By treating a nonprofit library like a commercial scraper, publishers are penalizing an institution that is actively working to mitigate the very problem they fear.

Journalists Are Losing Their Own Safety Net

Ironically, the entities most harmed by these restrictions are journalists themselves. The Wayback Machine is frequently used by reporters to verify past statements, recover deleted tweets, and track changes in government documents. Mark Graham, director of the Wayback Machine, reported that senior researchers at The New York Times have thanked him for the archive’s utility, admitting that Times staffers rely on it to find material that no longer exists on the live web. Without this independent steward, the ability to hold powerful institutions accountable through empirical historical records is significantly weakened.

The Scale of Digital Decay

The necessity of web archiving is underscored by the fragility of online content. A 2024 Pew Research Center study tracking one million webpages over a decade found that one-quarter of those pages were inaccessible by the end of the study, with a 38 percent loss rate for pages existing in 2013. The Internet Archive currently preserves about 16 percent of these vanished URLs. This includes the archives of defunct publications like the Rocky Mountain News, which won four Pulitzer Prizes before shutting down in 2015. Without the Wayback Machine, such work becomes irretrievable, leaving holes in the historical record that cannot be filled by current news organizations alone.

Key Takeaways

  • Over 340 local-news sites have restricted Internet Archive access as of May 2026.
  • The New York Times claims the Wayback Machine provides "unfettered access" to its work, bypassing paywalls.
  • Pew Research data shows a 25 percent inaccessibility rate for webpages tracked over a ten-year period.
  • Journalists, including those at The New York Times, actively rely on the Wayback Machine for fact-checking and historical verification.

The Bottom Line

Publishers are firing at the wrong target; blocking the Internet Archive doesn’t stop AI companies, who have other data sources, but it does destroy the only reliable, independent audit trail of the web. We are trading long-term historical integrity for a false sense of security in the fight against machine learning.