“The survival of the arts depends on technology, but not in the way most funders think." In a recent article, Koven Smith, technologist and former Senior Director for Arts at Knight Foundation, states that over the years funders have prioritized short-term technological innovation at the cost of stability. That has led to an infrastructure crisis in cultural institutions accumulating large amounts of technical debt. The Overview spoke with Smith about what it would take to fix it.
What lies at the root of the technology crisis in cultural institutions?
Koven Smith: In the arts sector technology has never become part of normal practice. For most arts leaders, and especially for funders, any type of technology still feels inherently innovative. That makes it hard for them to distinguish between technology that helps their organization survive or technology for experimental exhibitions without a sustainable outcome.
Josh Goldblum, founder of experience design agency Bluecadet, mentioned in a comment on LinkedIn that in larger institutions, smart technology staff members can "Trojan horse" some permanent infrastructure into an innovation project. But most arts organizations lack the expertise to do that. As a funder who designed technology grants, I often dealt directly with the museum directors who didn’t necessarily understand what they were signing up for.
So it’s also a communication problem between funders and museum leaders because they miss the technological expertise to clearly express their funding needs?
The institution has to deliver arguments for its technology-funding needs, and that is challenging, because funders often don't have the technological knowledge either. Neither party can communicate effectively with each other. Part of the issue is that Art + Tech doesn’t have a natural home, in either traditional art funding or emerging technology funding.
Some foundations specializing in technology financing have no idea how serious the infrastructure problem is for the cultural sector. They believe that museums are simply not open to new technologies, which is a misconception.
"Funders should approach the technical debt problem first and foremost as a staffing problem." - Koven Smith
You have worked on both sides, in the funding space and in cultural institutions, so you have a complete view of the entire problem. From your experience, where do you see the most pressing infrastructure issues in the arts sector?
That depends on the domain. As a funder, I worked across many different cultural categories, from the performing to the visual arts, and with many museums. Most of them struggle with collection management infrastructure, but even basic projects like website redesigns require outside funding and only happen very rarely.
And that’s mainly because of a lack of in-house expertise, which is the biggest infrastructure gap at cultural organizations. So I think funders should approach the technical debt problem first and foremost as a staffing problem. You can also see that gap in the sector’s incoherent response to AI.
"Museums address the AI challenge individually when what they really need is collective bargaining power. "
– Koven Smith
In which way?
Museums address the AI challenge individually when what they really need is collective bargaining power. Take collection management, for example: There's no common infrastructure nor a policy around licensing their collections as AI training data, even though this is exactly what AI companies are scraping. Without frameworks in place, the bot traffic has gotten out of hand. Administrators of museum websites tell me they get constantly hit by AI scrapers causing collection sites to crash.
This has compounding effects, with some museums pulling their collections back from public view. There are initiatives like Creative Commons’ CC Signals project that provide frameworks to help content owners express how and if they want their works used in AI training. But without collective infrastructure, museums can't effectively participate in initiatives like this or agree on how their data should be used.
Are museums effectively powerless against AI companies?
Larger institutions like The Met or The Getty have some bargaining power, but a small to mid-sized museum doesn’t. I still remember the start of the Google Arts and Culture project, which featured content from over 2000 museums and archives that partnered with the company. The larger institutions were able to negotiate changes to the contract between them and Google. The museum I worked for at the time was not in that position; it was basically a take-it-or-leave-it situation.
However, if all the art museums in the United States had an organization that represents their interests, then they could collectively determine under which conditions AI companies are allowed to use their collections as training data.
It will be interesting, assuming issues around training data and scraping can be resolved, to see what the utility of LLMs will be in the museum space. Aaron Cope, Lead Developer at the SFO museum, has been doing some interesting work there. He prompted ChatGPT several times to see what it would answer about the museum and its collections. The outputs were wildly unpredictable and in the museum space this is not an outcome we can accept.
"Museums have to make sure that their visitor data is not exploited the way consumer data is on platforms like Amazon." - Koven Smith
That's not only the case with AI. Other technologies that museums purchase from third-party providers can also erode trust in institutions.
Cultural institutions must understand the terms of service and privacy policies they inherit when choosing tools like ticketing systems. Museums have to make sure that their visitor data is not exploited the way consumer data is on platforms like Amazon. Trust is a museum’s most important currency. Once that trust is eroded, they’ll never get it back.
The problem is that arts organizations don’t have the resources to build their own systems from scratch; that’s just not realistic or practical. We need tools that provide utility without increasing privacy risks for visitors. And, right now, especially many small- to mid-sized arts institutions don’t have the expertise to make that decision in an informed way.
Everything leads back to the shortage of skilled workers. What funding initiatives exist that address the expertise problem and how do they work?
Shortly before I joined Knight Foundation, there was an initiative to fund individual technology positions at museums. It was an experiment to see what would happen if, instead of funding some flashy technology, they funded technologists. I believe this was the right move, but it wasn’t a cure-all.
While a few of the positions worked out and remain in the sector, several of the technologists started at institutions with decaying infrastructure–everybody worked on ancient desktops and outdated software. They were hired with the expectation they could leapfrog their organizations directly into digital transformation. And unfortunately, almost all of those staff members ended up leaving, because the expectations were absolutely unreasonable with the infrastructure that was in place.
Are there approaches that could solve this in a resource-friendly way?
Similar to the AI challenge, I’m imagining a collective solution. For example, every museum needs an information architect. However, most institutions don’t require that as a full-time position, especially not small or mid-sized museums. A larger service organization that provides this expertise across the arts sector could help to vet the technology that gets used. For example, it could represent all museum websites from a given region, making sure that they run with the latest technology and handle migrations from one platform to another if needed. The downside with one service provider designing sites for hundreds of museums is that they’ll all look the same. However, those sites will be up and running and at far less cost.
Of course you’ll also need some in-house expertise. But even with resources to hire, museums often face a pipeline problem for technical staff. I remember an institution I worked with in Detroit that ramped up their digital archiving program, and even with funding it took them a year to find an archivist.
"The Knight Art + Tech Expansion Fund seeks to support artists and arts organizations in Knight communities by equipping them with the tools they need to expand the use of technology in their work and engage audiences in innovative ways."
What type of funding initiatives currently exist in the arts sector that address the technical debt issue?
I see real potential in the Doris Duke Technology Lab and Bloomberg Accelerator models, which focus on technology adaptation over simple innovation. During my work at the Knight Foundation, we also ran an initiative called the Art + Tech Expansion Fund supporting arts organizations and individual artists. We asked applicants to address expertise or technology needs where one-time funding could increase their capacity over a long stretch of time.
Initiatives like these aren't very common in the funding space. Normally, open call programs are project-based, especially in the arts. And I’d love to see more funders pick up on that model, which explicitly addresses long-term technical problems.
You mentioned the Art + Tech Expansion Fund as a rather unusual project in the funding landscape. What was the most unconventional application you received?
One application I'll never forget came from an oboist, who was hired to do a lot of outdoor performances, but her sheet music kept blowing away in the wind. Picture her at a wedding, halfway through a piece, a gust of wind scattering her sheet music across the lawn or turning pages halfway through.
In her application she stated: “If I had the means to transition all of my sheet music to an iPad, then I could play in low light and outdoor situations and take on more gigs.” It was such a small request that had such a big impact on how many performances per year she could do. When we designed the program we couldn’t necessarily anticipate those types of use cases. Another memorable experience was a trip to Charlotte, the first city Knight supported with the Art + Tech Expansion Fund. There we had a meeting with all of the grantees who had received financial support through the fund. Some told us: “We redesigned our website and now we can finally do online ticketing, which increased our sales.” Receiving that direct feedback was incredibly rewarding.
