How An Archeologist Used AI To Uncover Treasures From Bob Dylan And More

This blog first appeared as Steve Wunker's piece for Forbes

​Archeology isn’t just for uncovering relics of ancient civilizations. What if some of the principles of archeology could help locate modern-day media treasures, match them to current needs, and give them new life? And what if AI could make it happen?

That was the bold premise of trained archeologist and musician Peter Agelasto, who teamed with photographer and entrepreneur Richard Averitt to create a company called Starchive. The duo knew that artists of all stripes can have huge catalogs – photos, tapes, and more – that often get largely forgotten. Like old artifacts, however, they can be tagged to find them easily, set in their proper contexts, and refer to when the need arises.

Unearthing Dylan, Terkel, and More

The company got started by working with the Bob Dylan Music Company. Dylan had left thousands of artifacts, like photos from old tours, but there was no easy way to find what was relevant for contemporary uses. Starchive created a way. Then they brought that capability to other types of creators. The broadcaster and historian Studs Terkel, for instance, had left over 9,000 hours of recordings in a Chicago radio archive. With the help of AI partner HyperNatural, Starchive could “listen” to the recordings, determine what content meant, and relate that to users searches or trending topics in their socials.

Generative AI opened the floodgates. Now, the AI could match meanings even if the tags weren’t using precisely the keyword that the user typed into the prompt. This could vastly expand the set of items found.

As the company progressed, it understood more about the magnitude of media artifacts. Agelasto says,
​

“It used to take hundreds or thousands of years to bury artifacts, but modern digital artifacts are buried nearly instantaneously - think of an Instagram thread. What if AI could help excavate them and understand them to make disorganization a relic of the past?”
​

The world abounds with content indexing tools, like Google Photos. How was Starchive going to be different? The team sought to understand the Jobs to be Done of its users. Averitt explains,
​

“It used to be that the cost of production and distribution meant only the Dylans of the world created assets with enough commercial value that it was worth the effort to archive them, but we live in an era of niche audiences with insatiable content demands. With AI as a supporting player, every creator with a following can maximize their opportunity and content in ways never before possible.”

Finding Business Treasures

The company has since found that artists aren’t the only ones with old content waiting to be re-discovered. Writers, social media influencers, B2B marketers, and others have produced huge quantities of content over the years. What if AI could help not just discover old content in response to prompts, but match content to particular customers and automatically send them old content that’s selected for them? What if AI could even do that in a super-timely way, such as in response to a social media post someone makes, so humans didn’t need to be involved after the content is first produced? That’s a use case that Starchive is actively prototyping today.

The company also sees AI impacting the content itself. Averitt says,

“Scale used to be a thing that only major enterprises had. But, today, individual content creators are becoming brands and creating content at scale. I have one friend, for instance, who created 100,000 new images in three months using the AI tool Stable Diffusion. These people need tools to find the right content at the right time as the amount of content simply explodes.”

​AI doesn’t need to generate only new content. Making AI a tool of content archeology, this company is leveraging new capabilities to repurpose content from yesterday as well as yesteryear. No AI tool is going to generate Bob Dylan’s type of content, but it can give his old treasures new life.

By Steve Wunker