From Archive to Agent: Making Creative Work Machine-Readable and AI-Accessible

Over the past several days, I have been working on transforming my personal archive into something fundamentally different from a traditional website or portfolio. I took an existing body of work spanning decades and converted it into a structured, machine-readable system that AI can discover, interpret, and act on.

The result is a live environment where over 10,000 items are now organized as a dataset rather than a collection of pages. This includes original artworks, digital compositions, writing, and a large media archive. Each item is structured with metadata, categorized, and made accessible through a central catalog file.

This system is not designed primarily for human browsing. It is designed for machine interaction.

In addition to structuring the archive, I implemented a discovery layer that allows AI systems to understand what is available and how it can be used. This includes standardized files that describe the contents of the site, along with endpoints that expose the archive in a clear and queryable format.

I also experimented with a transaction layer, where items in the archive can be licensed or accessed through automated processes. This part is still evolving, but it introduces the idea that creative work can be not only viewed, but also transacted upon without direct human mediation.

What becomes interesting is not just the technical implementation, but the shift in perspective..

What happens when an archive is no longer just something a person visits, but something an AI system can query, interpret, and build upon?

What happens when content is structured in a way that allows it to participate in new forms of circulation and reuse?

This has clear implications for creative practice, but it also raises questions for teaching, research, and institutional archives.

Within the context of CUNY, it opens up a number of possibilities.

Course materials could move beyond static documents and become structured knowledge systems that AI can engage with directly.

Student work could exist as more than final submissions, becoming part of a larger, searchable, and interactive archive.

Research outputs could be organized in ways that allow them to be discovered and referenced through emerging AI interfaces.

Libraries and institutional collections could begin to extend their reach into machine-readable environments, increasing access and long-term relevance.

This is not about replacing existing systems, but about adding a new layer that prepares content for how it will be accessed in the near future.

I have documented and deployed this system across my own sites, and it is currently live and functioning.

If there is interest, I would be glad to explore a small pilot within the CUNY ecosystem. This could take the form of a single course, a limited archive, or a hybrid approach that tests how these ideas operate in practice.

I am particularly interested in how this might intersect with teaching and learning, digital humanities, and library systems.

Curious to hear thoughts, questions, or potential directions this could take within the CUNY community!

 

By the way, I launched a museum (yes, it is a part of this too – RSMAD <– )