A new paper submitted to arXiv by Peter Williams, a researcher at the Center for Astrophysics | Harvard & Smithsonian, argues that artificial intelligence is providing the necessary tools to execute a large-scale 'enclosure' of the scientific information commons. Submitted on October 19, 2026, the work posits that the scientific literature, traditionally viewed as an intangible resource shared by everyone, is facing a threat analogous to the historical enclosure of agricultural lands in preindustrial England.

Historical Parallels and Economic Frameworks

Williams grounds his argument in economic and political frameworks derived from the history of common lands. He distinguishes the fate of the scientific commons from the 'tragedy of the commons' theory, which suggests resources are depleted by overuse. Instead, he points to the enclosure movement of the 17th and 18th centuries, where common lands were transformed into private property. The paper notes that many historical observers viewed this transformation as theft at a societal scale, a framing the author applies to the current technological shift.

AI as the Tool for Privatization

The core technical argument is that AI enables a previously infeasible mechanism for enclosing information. Unlike physical land, the information commons is robust and immensely valuable, often ranking among humanity's greatest achievements. However, Williams argues that AI tools allow for the conversion of this shared resource into private wealth. This process does not merely involve scraping data; it actively undermines the social systems that support open science through the proliferation of 'slop spam,' which dilutes the value of the commons for genuine contributors.

Implications for the Scientific Community

The paper serves as a primer on these frameworks for researchers and developers who may feel uneasy about the current trajectory of AI in academia. Williams suggests that the 'foreboding feeling' many scientists have regarding AI is justified, framing the technology not just as a disruptive tool but as an attack on a way of life. By providing a scholarly basis for these concerns, the work invites the community to interrogate the structural changes being wrought by large language models and automated research agents.

Key Takeaways

  • The scientific literature is analyzed as an 'information commons' akin to historical agricultural lands.
  • AI provides the technical means to enclose this commons, converting shared knowledge into private wealth.
  • The 'tragedy of the commons' is not the primary threat; rather, it is the active enclosure movement enabled by AI.
  • 'Slop spam' generated by AI tools undermines the social systems supporting open scientific exchange.

The Bottom Line

For those of us building on top of open datasets, this paper is a stark reminder that our foundational infrastructure is not guaranteed to remain public. If AI continues to function as a mechanism for enclosure, the open web as we know it may become a relic, much like the common lands of 17th-century England.