By Douglas McLennan, Editor, ArtsJournal.com

It seems like such a straightforward and valuable endeavor, measuring participation in the arts. Count tickets sold; ask what people are reading, watching, or listening to; ask whether and what kind of art people make; find out what kind of arts activities they’ve participated in. The National Endowment for the Arts’ every-five-years (or so) report on participation in the arts is a noble attempt to capture trends in both the interest of Americans in the arts and also what form that interest takes.

The top-line trend in the 2022 data indicates that arts participation over the past few decades has stayed more or less steady, with about half the U.S. population reporting that they created and/or performed art that year. Three quarters of U.S. adults consumed art via media in 2022. It also attempts to show shifts in how people participate; for example, online participation has grown significantly while traditional arts such as classical music, ballet, and theatre have declined. Those who work in the arts can feel the chill in the climate; the NEA tries to put some numbers to it.

But “participation” has always seemed something of a slippery concept. If I’m making pottery or playing the piano, I’m certainly participating in an active way, developing skills, making something, learning. How about if I go to the theater to watch a play? Sure, I’m not “making” theater, but if you believe that an audience is a participant in a creative act, then that works. By that measure, if I attend an orchestra concert, I count as a participant. But what if I listen to Gustav Mahler’s Symphony No. 2 on my speakers at home? I intentionally sought out the performance, so maybe that seems like participating. How about if I have a classical music stream going on in the background during the day, but I’m not actively listening? Or how about if I spend my day every day online reading articles but don’t read books? Have I stopped participating in reading, or are we making a value judgment about what the arts are or what is worthwhile participatory activity?

Then there’s the matter of density. Some forms of skill-based participation require continual practice to amount to anything of significance. Other one-of-a-kind experiences are so indelible that they need not be repeated to have lasting impact. So, can we generically determine what the threshold for an arts experience is to deserve being counted, or whether a single art experience in the past year means anything at all?

The average American spends around 4.5 hours a day on their smartphones, with about half of that on social media. Teenagers spend significantly more—about 8 hours and 39 minutes, according to one recent study. Most of that is on social networks such as TikTok, Instagram, and YouTube. This activity is certainly participatory, often highly creative, and interactive with both the content and collaborative with those who created it. Is this participation in the arts, and if so, how do you measure it?

Over the past 20 years, notions of who is an artist and what is an audience have changed significantly. Pre-internet, the lines were pretty clear about the binary relationship between artist and audience. Artists created and audience consumed. In today’s digital world, the landscape is fluid—we create and express our identities by what we choose to share online. Sharing, or curating what we encounter both online and in the real world, is perceived as a creative act. In the online world, art doesn’t become activated until people decide to “do” something measurable with it (share/like/comment/subscribe). Relationships between audiences and artists have become multidimensional. Sometimes you’re the artist, sometimes you’re the audience, and sometimes, in a shared environment, you’re both. Technology has put creative tools in the hands of billions, and both the amount of content being made and its accessibility have exponentially increased. How to measure this participation?

Popularity has always been an imperfect way of measuring the value of art. Yet pre-internet, popularity—albums or books sold, television or radio ratings, concert or theater ticket sales—was a credible way of counting what resonated in the culture. In that model, sales correlated with popularity because buying physical product or live experience required the consumer to spend something, validating the choice.

In the era of digital reproduction where copying and distribution are essentially free, clicking to watch or listen in an endless stream costs little commitment, and anything that fails to hold attention is quickly banished for the next thing. Value is now arguably detached from popularity. But the measure for successful content has remained attention, even in its devalued state, causing distortions of that value. Incendiary content, for example, demands a response, leading to boosts in producing such content over other, possibly more thoughtful, impactful content.

In an economy where attention has become the primary currency, and where users are incentivized to create their own content to gain followers and feed egos as “influencers,” the sheer amount of content and sharing has exploded. It’s estimated that 328.77 million terabytes of content are uploaded to the internet every day. Spotify reports 120,000 new tracks uploaded to its platform every day. YouTube users upload 3.7 million new videos (271,330 hours) every day.

In contrast to 20th-century popular culture, where mass media television ratings were measured in the mere millions, now hundreds of thousands of videos and music tracks get more than 100 million listeners/viewers, with a growing number exceeding a billion views. But what does it mean, practically, to reach a billion views versus 100 million? What’s the participation measure?

The flood of content easily overwhelms users’ ability to sort through it, and digital platforms have developed algorithms to organize and recommend. Whereas in the early days of the web, content-finding was a search (pull) activity, today it is a receive (push) experience. Some 70 percent of the videos watched on YouTube are chosen by algorithms rather than users having looked for them. Spotify reports that a third of the tracks played from its 100 million-track database are chosen by algorithm. If I didn’t choose what I saw or heard, am I still participating in it?

Music and video subscription services bundle up enormous catalogs of product, shifting the mindset of the consumer from relationships with artists to relationships with content catalogs, meaning that new artists have an increasingly difficult time breaking through. And because the streaming model pays out to artists on a popularity model—per-stream compensation—the model doesn’t care whether you’re Taylor Swift or a track of white noise in the form of the sound of a vacuum cleaner.

Indeed, fake music, “functional” music, and AI-generated music have become an enormous problem, as fake books uploaded to Amazon’s store are created to game the company’s measurement/reward system. The French music-streaming company Deezer reports its catalog recently grew to 200 million tracks, but discovered that at least 25 million of them were fraudulent before removing them.

This has led to music streamers reforming their popularity algorithm models. Quantity has overwhelmed quality content and its ability to be found. Spotify and Deezer both recently began halting payouts for tracks that have fewer than 1,000 plays. Deezer also now pays more for tracks that have been sought out by the listener rather than for those that are served as part of an algorithmic stream. Reforms in the model stem from two factors: musicians complaining that it has become tougher to earn money from their work; and a recognition that just as all content is not equal, neither are listens or views.

The music industry calls its new model artist-centric—attempting to reward artists who have committed fan bases—versus user-centric, an older model that pays based on dividing total revenues by share of plays. The first cracks in the tyranny of the social web, which opaquely chooses winners and losers in the attention algorithm wars, may be starting to show. There is increasing evidence that the social web is in advanced decay for both audience and creators, and that the participation models being rewarded online are deeply—even injuriously—flawed.

The NEA’s participation measurements are weighted toward nonprofit, traditional art forms, and it could be argued that online creative participation, while still growing, is a different animal from traditional arts participation. But online participation models, which initially seemed to promise greater access to the arts generally (as did the invention of television some 75 years before it), have significantly helped to hollow out real-world participation and the infrastructure that supports real-world arts production.

Into this landscape comes generative artificial intelligence. If the social web of Web 2.0 was about curation of content, Web 3.0 is about the exponential creation of new content. On the one hand, AI chatbots offer the potential to wipe away technical barriers to creativity. No need to learn how to read music or orchestrate when you can describe what you want in detail, and the machine will create finished symphonies—and at mass scale. These opportunities might unleash new human creativity at a massive scale, even making possible new art forms.

On the other hand, the flood of creative content we’re already navigating will massively expand, optimized to being found and served to users by algorithms. Those algorithms in turn will train on AI-generated work, which will further optimize in increasingly tighter loops. And what about human artists whose work sits even slightly outside algorithms? If not rendered invisible, they’ll be relegated to minor channels that will make it difficult to find an audience or find support.

But there’s a possible other scenario—one in which the tyranny of the platforms’ algorithmic power to determine winners and losers is broken and replaced by personal smart AI agents able to connect artists directly with fans.

Any of these AI scenarios could be wildly off the mark. But I suggest them in recognition that measuring participation in the arts has, in the digital age, become a multidimensional and ever-shifting puzzle. Attention, engagement, and participation are now shaped and controlled by unseen forces far beyond the scope and imagination of the participation models the original NEA studies sought to measure. The AI-driven Synthetic Age around the corner could result in a vast wasteland of minimally engaged bored automatons. Or it might fuel new hunger for human-built, real-world physical art experiences that affect us deeply. In any case, the definition of participating in an artistic experience will most certainly continue to evolve.