Performances and exhibitions have long been confined to the realm of "feeling" and "atmosphere." Audience reactions to performances and exhibitions were predictable only through the volume of applause, post-show surveys, and videos and reviews posted on social media. However, digital transformation and advancements in AI technology are disrupting this assumption. Now, even in the arts industry, audiences are no longer abstract entities, but are being redefined as data subjects whose actions and reactions can be interpreted.

Market Needs: It's Not About Audience Size, It's About Audience Understanding

The core challenge facing the performing arts and exhibition industry isn't simply increasing audience numbers. Understanding which audiences respond, when, and to what within the same budget and space has become crucial.

• Why do certain performances have such high repeat attendance rates?
• Where in the exhibition route does dwell time decrease dramatically?
• When does a particular production element diminish audience immersion?

These questions are difficult to answer with traditional qualitative feedback alone. This is where AI-based audience analysis becomes necessary.

Approaching the Problem: Interpreting Audience Behavior in "Real-Time Context"

A key use of AI in performances and exhibitions is the patterning and contextualization of audience behavior. Rather than simply assessing "saw/didn't see," it comprehensively analyzes audience movements, gaze, and reaction speed.

Representative examples are as follows:

1. Audience Reaction Analysis: AI analyzes seat-by-seat eye tracking, applause timing, and audience movement data to determine the level of immersion in each scene. This allows directors to adjust the duration or lighting of specific scenes, or even redesign the rhythm of the performance itself.
2. Analysis of exhibition space movement and dwell time: AI vision technology and sensors are used to analyze visitor movement paths and dwell time. If visitors spend a short time in front of a particular artwork, the data can reveal potential issues with the artwork's presentation, lighting, or layout.
3. Audience segmentation and content optimization: Audience segmentation is achieved by synthesizing factors such as visit time, viewing speed, and repeat visits, and then incorporating these into future exhibition and performance marketing strategies. This allows for audience-focused design from the planning stage, rather than simply promoting the event.

Challenges from a technology, design, and security perspective

While AI audience analytics is fascinating, the challenges it presents are clear.

• Technical challenges: Algorithms that can accurately recognize objects in changing lighting conditions and crowded environments are needed.
• The challenge from a design perspective: the technology must operate as “invisibly” as possible so that the audience doesn’t feel watched.
• Security and Ethics Challenges: Facial recognition and behavioral data are directly linked to personal information. Data must be anonymized, its purpose must be clear, and its storage and use criteria must be transparent.

Iropke's thoughts: AI should be used as a solution to understand audiences.

Iropke doesn't view AI as a tool for "quantifying art." Rather, he sees it as a solution that helps performance and exhibition planners understand audiences more deeply. Audience analysis data doesn't detract from the artistic quality of a performance, but rather serves as a supplementary sensory aid, enabling creators to make more sophisticated choices. The key isn't the technology itself, but the questions it asks.

AI audience analysis in performances and exhibitions is not simply a trend. It represents a process in which the arts industry seeks a new balance between sensory and data, and the ability to design that balance will become a competitive edge in the future of performance and exhibition planning.