Pre-AI Content & The Steel Hoarders: A Parallel
- In a move to safeguard human expression,British software engineer and writer Graham-Cumming has launched a digital archive of pre-AI content.
- The website, lowbackgroundsteel.ai, serves as a repository for online resources untouched by AI-generated content.
- The archive features a range of sources, including a Wikipedia data dump from August 2022, project Gutenberg's collection of public domain books, and the Library of Congress photo...
Graham-Cumming spearheads a bold initiative: a pre-AI content archive designed to protect human creativity.This digital vault, mirroring the concept of low-background steel, collects and preserves online resources before the pervasive influence of artificial intelligence. Discover how this archive houses diverse materials, including Wikipedia data dumps, Project Gutenberg books, and the Arctic Code Vault. The project aims to chronicle human expression and offers a crucial perspective on our evolving interaction with technology, safeguarding human-generated content for future historians. Graham-Cumming’s actions are a testament to human knowledge’s value.News Directory 3 is following this story with intense interest. Examine the parallels and contemplate the future of creative endeavors, where the pre-AI historical context and the role it played has been catalogued.Discover what’s next …
graham-Cumming Creates archive to Preserve Human Creativity in age of AI
Updated June 24,2025
In a move to safeguard human expression,British software engineer and writer Graham-Cumming has launched a digital archive of pre-AI content. Known for his tech preservation efforts, including his successful petition for a government apology to codebreaker Alan Turing, Graham-Cumming’s latest project aims to document human creativity before the rise of artificial intelligence.
The website, lowbackgroundsteel.ai, serves as a repository for online resources untouched by AI-generated content. graham-cumming said he created the site in March 2023 as a “clearinghouse” for these materials.
The archive features a range of sources, including a Wikipedia data dump from August 2022, project Gutenberg’s collection of public domain books, and the Library of Congress photo archive. Also included is GitHub’s Arctic Code Vault, a collection of open-source code buried in a former coal mine near the North pole in February 2020. The wordfreq project, frozen before AI contamination, also appears.
The site accepts submissions through its Tumblr page. Graham-Cumming stresses the project’s goal is to document human creativity, not to make an anti-AI statement. He drew a parallel to low-background steel, which became less essential after atmospheric nuclear testing ended. Whether pre-AI content will follow a similar path remains to be seen.
Protecting sources of human creativity, including archival ones, may prove valuable, Graham-Cumming said.He proposed creating a “cryptographic ark” in 2020-a timestamped archive of pre-AI media that future historians could verify as authentic. He added that AI could cloud the historical record.
For now, lowbackgroundsteel.ai offers a catalog of human expression from what may be considered the last pre-AI era.It marks the boundary between human-generated and hybrid human-AI cultures, potentially aiding in understanding how human interaction evolved before AI’s widespread adoption.
What’s next
Graham-Cumming plans to continue expanding the archive, seeking further submissions to ensure a comprehensive collection of pre-AI human expression.