Chapter 1. A Theme for Museum Technology

“Humanizing the Digital” marked a bit of a rhetorical about-face. For most of its short history, museum technology has been about “digitizing the human.” Technologists endeavored to bring digital tools to bear on the mission of museums, in service of cultural heritage. At first that meant digitizing cultural collections, creating digital images and converting ledgers and card catalogues into databases. About 10 years ago, the process accelerated and gathered momentum, to include websites, social media, in-gallery interactives, and even museum processes and approaches. In the last two years, as the failings of digital culture have become shockingly apparent, we’ve seen a shift.

What if the most pressing problem today isn’t the lack of digital tools in cultural heritage, but rather the lack of humanity in digital culture? Are museum technologists in a unique position to critique and change digital culture?

These were the questions running through my mind as notifications from the MCN program committee flooded my phone. I’d been reflecting on the impact of Pink Art, an exhibition that had recently closed at my museum, and deep in conversation with colleagues on the state of museum technology. When the program committee discussed ideas for a theme, I posted some quick thoughts on “Humanizing the Digital,” and it stuck. Once it’s official, a good conference theme means many things to many people; I’m grateful for the chance to think more about what it meant to me, and how it came to be.

For a mere tagline, the conference theme actually encompasses a fair bit of history, idealism gone astray, and newfound opportunity.

1.

For me, Pink Art suggested a new direction for museum technology. Pink Art was an exhibition I helped develop at the Williams College Museum of Art that explored WCMA’s collection through the lens of color. From the start, the exhibition was a “digitizing humanity” project, by the end, it was clear that the lasting impact was in “humanizing the digital.”

In order to look at fifteen thousand objects through “rose-colored glasses” (so to speak), we leveraged digital tools and approaches. We had a huge head start, thanks to all the labor that went into creating and keeping the digital images and records we required. (Accession Number, an earlier WCMA exhibition, shone a light on the history of WCMA’s collection data, gathered from the first curator’s leather-bound ledger, then painstakingly transcribed to index cards, to a database, and on to another database.) All that was a prelude to a recent digital initiative called WCMA Digital that was working to augment, package and share WCMA’s collection data and images, making them available for research and creative use; this set the stage for Pink Art.

With data and images at the ready, how then do you rank a collection by “pinkness”? You start by defining the color “pink.” It turns out everyone has a different idea of what pink is, and no one on the exhibition team—not the curators, not the exhibition designer, not interpretation or technologists—wanted to decide for everyone. And so, we crowdsourced the definition of “pink.” A simple web app was developed that encouraged users to select the “pink” squares from a 4x4 grid of colors. Submit, select, repeat. The app provided a somewhat soothing, rather mindless, sometimes surprisingly social activity; in return, with each submission, the app was refining a model that defined “pink.”

Now that we had a definition of “pink,” we could rank the collection images. It turns out that defining what makes a painting “pink” is nearly as hard as defining the color “pink.” Is it the picture whose digital image has the most pixels closest to “pink”? What about the painting whose small patch of “pink” nevertheless utterly defined the work of art? Was a tiny pink picture pinker than a giant picture with a fraction of pink? What about photographs that turn pink with age? Five computer science students crafted algorithms to give a “pink”-ness score to each collection image, and to show their work. The algorithms were as individual as the students: “Islandize,” for instance, prefered art with solid patches of pink; “Toon” started by transforming images to pictures akin to an animation cell (because the author loves anime); “Crayola” matched colors to one of eleven crayon ”pinks.” There was no right approach—each algorithm provided its own very idiosyncratic perspective on the collection.

Each object now had not one pinkness score, but five. We combined them, averaged them, created color-coded spreadsheets and visualizations. The algorithms provided opinionated new perspectives to consider. The curatorial team consulted each, crafted a final checklist, and developed an installation and interpretation plan.

So, was Pink Art “curated by algorithms?” This was an important question we heard often. The easy answer was no, and the larger lesson was that no exhibition could ever be. Even if the data and images going into the algorithms were pristine (they’re textured and varied and incomplete), even if the algorithm was perfect and objective (like all algorithms, ours were handmade, full of human choices and biases), there would still need to be a human curator behind it all, caring for the data, directing the development of algorithms, deciding what to do with the output, arranging the works of art on the wall, and sharing the experience with the world.

While the exhibition offered a unique new perspective on the collection, connecting rarely-seen works to recent acquisitions through color, it was the algorithmic process that, for many, stole the show. The exhibition opened in the fall of 2017, at a time when media conversations about “fake news,” the excesses of innovation culture, and the dangers of data and algorithms were ascendent. The prevailing narrative espoused by the technology industry—that data and algorithms are objective, mechanical, unbiased and unquestionable—was cracking.

In public programs for Pink Art, conversations focused on the nature of algorithms. Our students wrote algorithms that were quirky, opinionated, expressive, and full of personality. The exhibition and the art drew out these qualities and made them more apparent. It’s plain to see that works of art can’t be reduced to data and pixels; they are physical objects that embody myriad human choices, decisions, indecisions; they are fluid and impossible to quantify; they bring a mythology of objectivity and mechanical certainty to its knees—even the myth that surrounds the words “data” and “algorithm.” Art helped us show that our data and algorithms were handmade, flawed, contingent, susceptible to bias, history, power; visitors learned the same is true for all algorithms. Further, the art objects also humanized the process of making. Working with art allowed the students to embrace the idea that making algorithms is a creative process, as much “creative writing” as “writing code.” It’s learning a language to express yourself in new ways.

We began the exhibition with a notion of showing off the power of digital tools to see the collection in new ways, and came out with important realizations of the imprint of humanity on digital processes and culture.

2.

Pink Art became part of an emerging conversation around “digital humanities” as well. Williams College is a premier liberal arts institution in which the arts are central; the art department has a prestigious history, and the art museum is a marquee public-facing institution known for innovative collaborations across campus. While there is no “digital humanities” program per se, there are faculty bringing digital tools and approaches to humanities research, and the art museum is well-positioned to connect the digital and the humanities through a project called WCMA Digital. The project enhances museum collection data, and encourages new digital approaches and tools. Indeed, the project even includes a Digital Humanities Postdoc Fellow to support faculty collaborations and projects. It makes perfect sense at Williams for the art museum to be the hub for innovation, collaboration and co-creation.

Bringing digital to the humanities was expected; the opportunity to bring the humanities to the digital was not. Enrollment in computer science classes has exploded at colleges and universities across the country, and there’s a shortage of professors—new Ph.D.’s are joining industry instead of academia. The Computer Science department at Williams faces these challenges as well, but it also sees an opportunity. Many institutions of higher learning—especially large research universities—can supply industry with proficient coders. As critiques of digital culture are quick to point out, however, the failings of industry are rarely technical. What industry needs are skilled technologists with a broader background—a perspective on how their work affects people, a grounding in history, social science, moral philosophy and creative practice. Williams is uniquely-suited to graduate humanist technologists like this, that can build mindfully and affect change in a post-digital culture.

At Williams, the art museum helps develop humanist technologists. WCMA’s collection data and image files provide a common language for interaction. It’s important that intro students can work with data quickly and easily; even beginners can download a csv and a zip file of collection thumbnails and start exploring right away. Moreover, museum data wears its heart on its sleeve: it’s quirky and textured, it reflects institutional history, it’s actively maintained and changing. It’s a great example of how data is subjective and contingent. Likewise, art images test the limits of taxonomy and analysis. WCMA Digital introduces students to the field of museum technology, and (with the Library and the IT department) to the notion of non-profit, non-commercial technology more broadly. Working with the museum expands the perception of coding as a creative process and personal practice, and code as an agent of change. The art museum provides a public interface for computer scientists to share ideas, to think through issues of usability and social impact.

3.

So does this paradigm shift from “digitizing humanity” to “humanizing digital”—the dynamic we saw in Pink Art, and in digital humanities on campus—have an analog in museum technology more broadly? Does it offer a chance to reflect on the past, and perhaps a path forward? This is the conversation I sought through MCN in 2018.

Ten years ago, museum technology was in an innovation phase, bringing an expanding array of new digital tools and approaches into the cultural sector. The web was mature and accessible, mobile was exploding, social media was transforming the landscape. Museum technologists weren’t alone in their idealism: powerful digital tools, expanded networks, more-connected people would mean a better society. Museums needed to act fast, or face disruption or irrelevance; this was a common driver for innovation culture. Much needed to be built, and fast: new mobile websites with content management systems, online collections, exhibition microsites, mobile apps, new visitor info screens, touchscreen interactives, new social media accounts, email templates. The new work required new processes, new approaches, a new mindset for organizations.

This innovation phase was largely additive; it didn’t fundamentally change the museum. Museum technologists benefitted from earlier work done in digitization and collections management; elaborate digital projects were built atop existing infrastructure without changing it or enhancing it. Indeed, in many cases the emerging “digital” positions in museums were separate from existing IT departments. This led to a “two-track I.T.” approach, one fast and charged with innovation, the other slow and responsible for maintenance. As in the larger innovation culture, speed and agility came at the cost of sustainability and lasting change.

Now we see where innovation culture failed. Social media networks became weaponized, data breaches undermined trust, trackers monetized attention, creative tools shuttered or transformed due to shareholder pressure. Even in museum technology, the idealism has faded. Innovative digital labs have disbanded or restructured. The ambitious digital projects built a decade ago have been retired; their impact is hard to discern. Meanwhile the systems software that was ignored in the rush to innovation—the collections management systems, the membership software, the DAMS—remains. It hasn’t changed at all from its pre-internet foundations. In some cases, platforms and approaches that were enlisted directly from industry now challenge museums’ core values (Facebook, Google Analytics); others that were enlisted as infrastructure have effectively disappeared (Flickr).

In some ways, the “digital humanities” approach, bringing digital tools to bear on the preservation and presentation of cultural heritage collections, has been transformative. (Digitization, for instance, has laid a foundation for the future.) In other ways, it’s not lived up to its promise.

What would it mean for the museum technology sector to embrace “humanizing the digital”? What would it look like for museums to offer an alternative to innovation culture and surveillance capitalism? How might museum technologists bring a humanities approach and values to bear on technology culture?

Here are a few ideas that came up in Denver, and percolated since:

1. Look long-term. Museums are charged with keeping objects forever; that gives museum technologists a unique and countercultural perspective. Museum technologists keep the objects’ stories alive for generations to come. Systems and formats will change, but the living, growing, changing data around collection objects—the human context that give the objects meaning—needs care and oversight.

2. Embrace maintenance. If innovation culture is about disruption, creative destruction, “moving fast and breaking things,” then maintenance culture (and museum culture) is about fixing, stewarding, sustaining. Perhaps ironically, maintenance encourages incremental improvement, and iterative, agile approaches.

3. Innovate on infrastructures. Innovation culture has given maintenance a bad name; infrastructure needs daring new ideas. Museum technologists are in a unique position to innovate on technology infrastructure, building sustainable core systems that make dynamic experiences easier to create and maintain. Technologists are already bringing new ideas to other museum infrastructures: processes, governance, org charts, staffing, strategy.

4. Focus on values. Technologies are not neutral. Each comes with a cost and values baked in. Museum technologists should strive to align technology decisions with museum values, and lead discussions around privacy, data stewardship, algorithmic bias, etc.

5. Design for social impact. Museum technologists work in public, in institutions that measure for public impact, in person and beyond. We have the opportunity to internalize institutional goals and build for good. Even internal museum tools and infrastructure can be designed for positive impact among staff.

How can a museum technologist change the world? There are precious few opportunities in the job market for creative technologists to work for good; the technology industry puts most to work in the service of surveillance capitalism. In museums (as in higher education, and libraries), technologists can build with an eye toward social good, sustainability, and long-term impact.

We face monumental challenges, and thoughtfully reorienting our relationship to technology is one. Is there a place for museums in finding solutions? Can museum technologists lead the way?