The Future of AI and Audience Engagement in the Arts

NY Live Arts hosted the Arts + AI Symposium, Saturday May 11, 2019. The Symposium was part of Live Arts, their annual humanities festival of arts and ideas. The 2019 Festival AI: Are You Brave Enough for The Brave New World? pondered a future with artificial intelligence, a technology that promises to revolutionize human existence. The festival headliner was a performance by discrete figures. Other activities included a hacking camp for teens and panel discussion on the Future of Work.

The Art + AI Symposium offered the sold-out crowd a speed dating style share-out from panel participants. As the manager in the group, I provided a perspective on AI focusing on how institutions will begin connecting the art to the audience using emerging technologies. The following article provides a summary of the frameworks and solutions I presented.


I am a practicing theatre artist and academic currently teaching at Carnegie Mellon University and running the Arts Management and Technology Lab. As such, I find myself at an intersection that aligns with my passions: I am obsessed with the potential of technology and the transformative power of art.


Sitting in the renaissance of web-based technology, it is important to remember that computers have been tools in theatre and arts management for decades.  I was in high school researching robotics when I first learned about Carnegie Mellon University’s role as a leader in AI and robots –-that was in the 1980s. A few years later, I kicked off my career in NYC as a production manager / stage manager who quickly found herself working as a managing director for a small off-Broadway theatre.  It was the early 1990s, and I was of the avant-garde with an in-house computer for contracts, spreadsheets for contact sheets and actively searching for a good (and affordable) computerized ticket system.  Of course, production systems for lighting and sound systems pre-dated our management tools.  By the late 1990s, I was working independently as a web site designer for institutions such as the early Shakespeare Globe Center (SGC) in London.  This was the early days of web 1.0, ftp and html.  My how things have advanced in the 20+ years.


It is critical, however, to remember that the arts are a humanist practice.  My experience with the SGC created a new understanding and fascination with how institutions manage their relationship with their audiences. In its early years the SGC was an academic experiment of sorts, learning through lived experience, similar to what performing and attending a show at the original Globe might have been like.  A critical question circulated around the capacity of the yard. Was there a magic number that could create a Globe Theatre yard experience that seemed to match accounts of the times?  Transformation of the audience through the artistic exchange is dependent on the audience’s lived experience. This experience is shaped by many things all critical to an audience’s ability to truly engage with art.


Using theatre as a metaphor, let me set the stage for the opportunities I see at play and poised for the future for AI + Art and audience engagement


Source: Author

Source: Author


The arts have the power to transform lives, bring us closer to our own humanity and offer understanding of the world around us as it is, has been, and could be. In our current system we have placed institutions firmly between the audience and the art.  Hence, it is the institutions responsibility to foster a connection to the art with the broadest available audience while also supporting all available modes for each audience member to find their connection and transformation.


In this system, then, institutions become the vehicle, the agent of change. In fact, that educative responsibility is why they are afforded not-for-profit status. 


In our current state, AI is or has the potential to be activated at each of these intersections.


The artist, like our esteemed host: Bill T. Jones and New York Live Arts and lead artist of discreet figures, Kyle MacDonald, can use forms of AI to open up the art to themselves in new ways in both the process and final product.  This often brings new people and/or skills into our production processes which potentially diversifies our work for even more audiences.


Institutions can and should be a partner in this process, whether producing or presenting performances or exhibitions.  Supporting artists investigating new technologies provides institutions with opportunities. Support might be as simple as writing grants to bring artists into an institution as part of a commission or a presenting series. Additionally, institutions can use AI to foster deeper avenues of connection  between the art and the audience. 


AI currently exists in our management tools.  We use of CRM and communication systems.   Particularly savvy communication departments are using machine learning tools to understand behaviors in their data. My interests and the conference focus on the lived experience of art in a shared space.  Location is a significant part of the arts system and one in which AI is increasingly playing a major role.  Be it a museum or theatre, space can be easily activated by AI.  Using beacons, RFID or other queuing systems, engagement between the art and the audience can be created to offer opportunities for engagement leading to deeper insight by the audience member. 


WolfBrown’s arc of engagement presents nodes of opportunities for engagement to consider as opportunities for AI interventions.


Source: Author’s Keynote slide representation of WolfBrown’s Arc of Engagement

Source: Author’s Keynote slide representation of WolfBrown’s Arc of Engagement


Everything in the arc that is controlled by the institution is, and should be, centered on deepening the experience between the audience and the art.   The ultimate goal is to deepen the impact echo.  That is what creates transformation (versus light, meaningless Instagram opportunities).


AI informed communication work can help in the early decision-making process, the build up, and even the post-processing stage. Yet, AI is having the most interesting opportunities for intervention in the artistic exchange.  For example, earlier this year, IBM partnered with the museum Pinacoteca de São Paulo. In this project, IBM Watson was trained to talk to people about the exhibit they were seeing. In this process, AI replaced the usual docent. A second IBM Watson intervention is perhaps more sublime:  Iris+

Source: IBM

Source: IBM


Iris+ was an intervention in a climate exhibit at the Museum of Tomorrow in Brazil . What made it unique was the use of the chatbot to ask the patron questions — not just deliver information.  As the audience went through the exhibit, they would ‘check-in’ with RFID chip cards and Iris would ask them  questions.  At the end of the exhibit they were shown how their answers fit into the rest of the audience and how their choices / thoughts might impact the world.


Screenshot 2019-06-27 13.07.13.png


Iris+ and other cohesive in-place audience engagement opportunities are in “closed systems.”  The system is controlled by the institution or the artist intervention such that it is baked into the experience. But what could this look like in an open system in a public space?


Last fall AMT Lab researchers embarked on a project entitled  Public Art +IoT = Civic Engagement that received an ideation grant from Metro21, a smart city institute. From a Smart City perspective, what made the project unique was the focus on quality of life improvement not cost-savings or increased convenience which are typical to smart city projects. 


As a public process, the project focused on a framework of values.  The intervention need to be participatory, interactive, reactive, communicative.

  • INCLUSIVE — anyone with any device / income / lifestyle / interest could participate

  • INSPIRING — it should offer MORE information and pathways than a text on a wall

  • SHAREABLE — social must be part if the plan

  • FEASIBLE So that OUR PARTNERS can continue.


Globally, cities and public art departments are creating unique technology solutions to forge connections between citizens and art.  However, most are either story-maps or location-based and require an app.  These solutions all required smart phones, knocking out the approximate 30% of the US population without a smart phone. 

Our solution was to offer audiences an opportunity to learn about a public art work, even ask questions about the art, during their in-place experience. The experiment used IBM Watson across both SMS and a mobile web site for 5 pieces of public art in downtown Pittsburgh. There were 5 chatbots, one for each art work, with prompts for participation coming via  social media geo-fenced ads, google maps or physical signs at each location with the SMS or QR codes


The location was chosen based upon its high density of potential participants with a cross section of lifestyle & incomes, as well as a density of potential public art. The artworks were chosen to see if different forms of art affected the engagement. The pieces spanned a ¾ mile length of downtown picture and included 1 temporary video installation, 1 mural, 2 statues and 1 living piece of art.

Source: Pittsburgh Art Places, from left to right: Yesterday’s Tomorrow, Streaming Sound, Agnes Katz Plaza, Liberty Avenue Musicians, Magnolias for Pittsburgh.

Source: Pittsburgh Art Places, from left to right: Yesterday’s Tomorrow, Streaming Sound, Agnes Katz Plaza, Liberty Avenue Musicians, Magnolias for Pittsburgh.


The core research questions for the experiment were:

  • Which paths did participants use?

  • What questions did public art audiences have? Did they exceed the curatorial / artist statement scope usually provided online?

  • What did the participant think about the experience of using a chatbot?


The experiment ran March 27 - April 27. Results were positive: people enjoy public art and want to engage or at least learn more about it.  Within the test model, the most popular pathway was fenced social media (Facebook and Snapchat specifically). A few people even engaged with each other on Facebook asynchronously discussing the art.  With over 1000 interactions, 10% came from Snapchat, 70% from Facebook and the remainder via SMS or the QR code.  


From research and a local experience perspective, it is clear that AI will have an impact on the practice of arts management and public programs. Be it a chatbot on a website, a robot functioning as docent, or a personalized delivery of contextual material via a mobile device, AI is, and will change audience’s experience and expectation for their level of interaction with art. AI is changing not only our art but the relationship our audience has with the art.  My hope for the future is that the solutions are designed for deepening the impact of the art beyond a really good instagram post.


Words of warning: It is important to remember as institutions get excited about the potential for AI that all change has potential unintended consequences.  We already know that the artificial intelligence in some of our media systems are not only biased but are creating less curious audiences increasingly accustomed to “more the same or similar.” This intellectual silo-ization echoes the bias and similar concerns around AI algorithms and social media, target marketing and fairly standardized communication strategies. Furthermore, there are concerns with accessibility. While technology can serve as a useful tool, personal technology choices are framed by economic opportunities that are skewed in the US.  Hence, if an institution wants to reach “everyone” or a majority then a clear analysis of local technology adoption and use must come into the conversation to prevent assumptions by those who may be more financially well-off or simply technologically inclined. Intentional use of technology with consideration of goals, audience and feasibility are all part of a smart technology plan.