AIC's 53rd Annual Meeting

How is machine learning used to create works of art? How do machine learning technologies work? What are the various software tools and programming languages that are available to artists? What are the conservation problems that arise with each of these techniques?

The software applications that artists use for creating works of art which integrate or are based on machine learning fall into several categories. For artists who do not know how to program, or prefer not to program, and/or do not have opportunities to collaborate with programmers, there are text-to-image applications in which an artist creates images generated through textual description. Examples from the New York City bitforms gallery exhibition DALL·E: Artificial Imagination (October 26–Dec 29, 2022) demonstrate this approach.[1] 

Newly created images using text-to-image techniques can be loosely based on predefined styles provided by the software authors or company; or the artist can “train” a model to use style-transfer based on the artist’s own original digital-born images or digital surrogates of physical artworks in order to instruct the software to computationally mimic the artist’s own or another style. The Whitney Museum of American Art’s xhairymutantx Embedding 2024 by Holly Herndon and Mat Dryhurst (2024) is an example of style-transfer.[2]

With advanced beginner or intermediate programming skills, artists may prefer writing original code such as Python scripts to generate new images based on style-transfer and other techniques. An artist at this level of programming skill can also write scripts to programmatically download images from the web that meet specific textual criteria, e.g., “watercolors of pink roses.” 

Building an original machine learning application requires great resources and advanced computational and programming skills. The artist Refik Anadol, in his talk at the Institute of Fine Arts in New York City on June 3, 2024, said that it can take over six months of teamwork at his studio to compile data and build the application for a work such as Unsupervised, which was exhibited at the Museum of Modern Art in New York City.[3] Managing a team to retrieve and prepare data sets, as well as collaborating with programmers to run and train machine learning models, requires extensive studio resources including hardware for data storage and processing, bespoke software that addresses the artist’s vision, and a staff with appropriate expertise. 

Each of these approaches brings up a specific set of questions regarding acquisition practices, documentation practices, preparation for future re-exhibition, and other conservation concerns. Answering these and other questions, focusing on the collaboration between institutions and collectors with artists and engineers, leads to conservation strategies for these fragile and complex artworks, as artists continue to explore the use of machine learning as an artistic medium.

[1] https://bitforms.art/exhibition/dall%C2%B7e-artificial-imagination/ 

[2] https://whitney.org/exhibitions/xhairymutantx

[3] https://www.moma.org/calendar/exhibitions/5535

As we consider new tools and technologies for working with the video signal, it can be interesting to look back at key periods of innovation for digital video editing and manipulation. Steina and Woody Vasulka are "pioneers" of video and new media art and technology who spent their careers exploring the innate potential of the signal and pushed for new tools to facilitate this exploration. Steina, in particular, was fascinated with advancements in real-time A/V processing for purposes of performances, interactivity, and immersive environments and worked with many engineers and software developers throughout the 90s to create and modify software for these purposes. In anticipation of an upcoming exhibit being organized by the MIT List Center, I have been going back into my days assisting the Vasulkas with their archive and exhuming rare videos documenting the development process for these tools which has led to further inquiry around what was not-yet possible to do with consumer-based open-source video tools from this time. This begs the question of what past efforts have been made to create artist-driven tools with an open-source ethos, the successes and failures of these efforts, and what archives of this content can do to better ensure these obscure and abstracted histories can be interwoven to form a more complete narrative around media histories.

In this paper, we present recent collaborative work at Tate to advance the understanding and care of artworks incorporating control systems. Artworks can make use of dynamic elements which require management and coordination; for example, the dimming of lights, the driving of motors, or the coordination of multiple channels of audio or video. At the heart of such artworks are control systems: sets of components, typically involving programmed computer hardware, which choreograph the sequence of actions desired by the artist. While many of these technologies overlap with those used in software and computer-based art—a medium that has been a focus of research at Tate over the past decade—they differ in their reduced emphasis on material specificity and their tendency to remain inconspicuous when the artwork is displayed. In light of these differences, we identified control systems as a distinct challenge that would benefit from further research.

Building on our experiences in the conservation of software-based art, we examined a range of artworks where control systems play a critical role. Reflecting on both commonalities and unique attributes, this investigation led to the development of guidance designed to assist conservators at Tate, including:

* guidelines for identifying control systems, including common component types, and how they differ from software-based artworks;

* key considerations when condition checking and documenting control systems;

* measures to prepare for the future translation of control system functionality to new technologies, as a response to obsolescence.

We found that many principles applied to the conservation of software-based art remained relevant, but the relative importance of these shifted and certain activities (such as disk imaging) were less useful. Our findings placed particular emphasis on understanding the control sequence—the series of actions enacted by the control system. This entails a variable process of analysis and documentation which may require specialist expertise and provides the key to migrating the control system to new technologies in the future.

We have adopted the “control system” label as a pragmatic means of highlighting the conservation challenges associated with a distinct yet diverse group of artworks. While this has helped us advance our understanding, it is clear that this grouping is not homogeneous and we encountered artworks that defy categorisation. This illustrates the limitations of medium-based terminology and the evolving nature of artistic practices which will continue to transcend medium-led conservation approaches. It underscores a need for well-resourced, interdisciplinary conservation work at points of acquisition and display, and for research time to be integrated into these processes as we continue to learn. With control systems present in many collections, and potentially falling under the radar of time-based media conservation projects, our findings have broader implications. We hope this paper will spark a wider conversation and foreground the power of interdisciplinary collaboration to influence future care and preservation strategies for these artworks.

Keywords: Control systems, time-based media

The iteration report, first proposed by Joanna Philips, has been used in conservation to document different manifestations of a time-based media artwork with the understanding that each iteration results in changes to the work. These reports often encourage the writer to reflect on the iteration in terms of the appearance of the work, decision-making processes that led to the final result, and to assess whether it was successful and why. 

In one instance at the Museum of Fine Arts, Houston, we used the iteration report on a loaned artwork to not only reflect on the final appearance of the artwork, but also to consider the entire installation process for time-based artworks at the museum. The Museum recently hired a time-based media conservator; thus, some of the aims of this iteration report were to diagnose inefficiencies in current workflows as well as creating an ideal iteration report to serve as a model for future reports. The report was written in a collaborative manner with stakeholders in the Audio/Visual and Registrarial Departments. We are choosing not to name the artwork because it is not owned by the museum. This artwork’s installation, which was its second iteration, necessitated purchasing new equipment and altering the exhibition space which resulted in delays to the opening. The installation also coincided with other installations with tight deadlines, which placed intense pressure on museum staff. The iteration report served as a vehicle that allowed us to track where workflows could be improved to avoid having those same challenges in the future.

To aid us in this reflexive practice, we made some modifications to the report.  We created a timeline to note every decision that was made about the work from the moment it was considered for exhibition through the end of the exhibition, expanding the focus beyond the final presentation of the work. We also rigorously documented the labor involved in installing the work and creating the report. After the report was written, it was used as a tool to inform workflows for installing other time-based media artworks. This use of the iteration report became a catalyst for change in the museum with respect to how time-based media artworks are understood and handled, and therefore became a tool for internal advocacy. This adaptation of the iteration report could serve as a model for other stewards who are advocating for improved time-based media workflows in their institutions. 

In this presentation, Emma Dickson and Cass Fino-Radin explore the critical role of interdisciplinary collaboration in advancing the field of time-based media art conservation. Aligning with the conference theme "What's Your Story: The Power of Collaborations," we argue that the most innovative and effective conservation practices emerge when we blur traditional role boundaries, deconstruct established hierarchies within our field, and facilitate opportunities to exchange practical skills between practitioners.

Through years of cross-disciplinary collaboration on the treatment and migration of complex and interactive time-based media artworks, we have developed and refined new methods for assessing treatments, expanding beyond traditional visual inspection and the limits of human perception. These new methods—which are reproducible and quite accessible—will be shared by illustrating their application to two specific works of art: Tall Ships (1992) by Gary Hill and Ten Thousand Cents (2008) by Aaron Koblin and Takashi Kawashima.

These case studies will illustrate the specific tactics and methodologies used to apply two new universal principles for assessing time-based media works that have emerged from our collaboration:

1. Automation of interactivity for consistent artwork testing

2. Measurable and time-synchronized comparison of iterations

As the field of time-based media conservation continues to mature, integrating these principles into practice is essential for maintaining the integrity of time-based media artworks through successive conservation treatments. By providing conservators with replicable, objective means of assessment, these techniques help minimize unintended alterations that would otherwise inevitably accumulate over time.

The development of these new methods demonstrates how interdisciplinarity when extended beyond collaboration into individual experience and training that bridges into one's collaborator's field, can enhance conservation practice and yield the kind of innovation our specialization needs to steward the art of today and tomorrow.

While larger museums move forward with their Time Based Media collections, smaller and mid-size institutions continue to struggle with preservation planning for these complex contemporary works without easy access to media conservators - especially if they are all digital. Maintenance Culture is a project created by Myriad, funded by the National Endowment for the Humanities, to address challenges related to preserving complex, born-digital, creative works in smaller institutions. From 2022 - 2024, Maintenance Culture brought together creators and maintainers of digital design, web art, time-based media art, virtual reality, and more to address pressing challenges of preserving these works in small institutions. Through a Design Charrette and various working groups, Myriad organized across institutions and disciplines to create events, workshops, and guidelines for creators and maintainers (curators, conservators, librarians, other cultural heritage workers) who preserve digital design, time-based media art, net art, augmented reality, and more. 

Workshops were offered in 6 cities across the U.S., focusing on mid-sized cities including Baltimore, Houston, Detroit, and New Orleans.  Participants included cultural heritage professionals from a wide range of institutions seeking to provide long-term access to complex digital creative works. Participants shared experiences, discussed best practices, and worked across disciplines to consider new solutions for preservation of complex objects.

Project staff had ideas about preserving born-digital works at the start of the project, but the addition of artists’ knowledge provided insights into their creative process, intentions, and skill sets that changed the course of the work. We will share information about the implementation of the project, insights learned through the project, ways that  collaborations with creators shaped the outcomes of the work, and lessons learned. We will include evaluation data showing workshop participants’ achievements and reactions, and we will also talk about the future of Maintenance Culture, which has secured a new round of funding from the NEH and will continue through at least 2026.