Flow acts as a technical term for the smoothness and continuity of the movement of goods, people, and information… When energy and resources, goods and data, crowds and individuals get to the right place at the right time in the right amount or number, without interruption, they seem to have a life of their own… In this way, the flow metaphor reveals the insidiousness of logistical power–as well as its ubiquity.

— Susan Zieger, Logistics and Power: Supply Chains from Slavery to Space

There were two moments during her most recent concert at The Bellwether in downtown Los Angeles when Santigold invited the crowd onto the stage with her. The first time, participants are instructed to follow the minimalist dance moves of her two dancers while they are graded by the audience via boos and cheers. The second moment – during her 2008 hit “Creator” — is a free for all. Dance how you want, she said. She only asked concertgoers put away their phones and experience the moment.

As Santigold weaved through her high-energy show, encouraging the equally high-energy attendees, I was struck by her ability to be both separate from the crowd and one with it. Her performance was about living in that moment. You were at the show and you were the show. We were all together. And we, she kept telling us, needed to be a force outside the show, in the world beyond.

She proclaimed her authorship, while acknowledging the participatory audience needed to create a successful show. “I see you singing along and I appreciate it,” she said, “I want you to really listen to the lyrics. I wrote them and they’re good and they’re important.”

She pointed at us, as we all sang the chorus from “Ushers of the New World.” She wanted us to acknowledge the import of the lyrics; we should take them in and use them later to power our actions and our hearts.

Much of the 2000s and 2010s discourse was about the concept of “visibility.” Social media opened new avenues for people to see and be seen. Power seemed to lie in who gets to tell their stories, in who gets to be made visible. It isn’t surprising that the controversial International Transgender Day of Visibility came about around 2010. (In recent years, many activists say visibility is not enough.)

As individuals sought visibility (and used it for marginal cultural gains), capital sought greater invisibility. Companies sought to make our interactions more opaque—filtering transactions through various app interfaces. The cab driver in his yellow cab became the gig worker in the ever-changing Uber or Lyft. Recently, this has gone a step further as the driver has become entirely invisible; allegedly driverless Waymos fill the streets. The delivery driver in her branded car became the anonymous “Dasher.” The building next door became a hotel overnight; you once thought it was a home.

In the face of a post-2020 reality, capital seeks to own both visibility and invisibility, primarily via AI. AI can help a user feel as if they are “creating” while regurgitating tired tropes mixed with thoughts stolen from other people. AI can mimic and, frequently, profits off mimicries of marginalized people. AI companies can clog our social media timelines with visible messages made by invisible hands.

AI makes less visible the work of all those who created it (whether or not they agreed to the creation in the first place). AI companies and other automation-focused tech companies obscure the necessity of large networks of data workers, content moderators, and other gig workers, furthering extractive relationships started with Western colonialism.

Even for those in the West, companies seek to use AI to obscure the workers needed to create work. Google recently announced they are moving to AI chatbots, rather than traditional search. This move will decimate the websites which create the search results. But it will also obscure the authorship of “who” is answering your query. They seek to hide the source of information in an algorithm. You have to trust they are correct. AI excels at authoritative answers, but not necessarily correct ones.

This invisibility sets the stage for tech companies to complete their global takeover. A recent study showed ChatGPT provided misinformation to users for 34% of the questions asked about the Scottish election. Is this election interference? Who can we blame? Media companies owned by investors who also invested in AI will likely give AI the “past exonerative tense” treatment once reserved for cops.

But, in the 2020s, invisibility has been a source of power for people, not just capital. Here in the United States, Tesla Takedown protests erupted after Donald Trump won his second term. A leaderless and faceless movement, they accomplished real wins as Tesla stock tumbled.

More recently, as video after video of graduates booing AI-boosting commencement speakers are released, we see the power of anonymous censure. It isn’t one student. It isn’t a movement. It is just a bunch of folks in one room who agree.

In both cases, it is the invisibility of those groups that ensures their safety and gives them their power. There is no “head” of Tesla Takedown able to be targeted by the state. There is no singular student who can have their diploma withheld as punishment, a coercive tactic of which universities have become fond.

Visibility, at the same time, becomes a threat used against capitalists. Zohran Mamdani, who is an actor of the state at the end of the day, pursues a populist image and is sometimes willing to stand in as a proxy for public sentiment. Billionaires melt down across various media platforms, whining because the mayor pointed out the building where venture capitalist Kenneth Griffin owns a $238 million dollar penthouse.

(This month, Griffin reversed course from his position as an AI skeptic and I wonder if this use of publicly available information had some impact on his about-face.)

As someone who spent a large part of her life doing standup, I hate crowd work clips. (I, of course, love doing crowd work. To paraphrase The Professional Joke Explainer, Myles Anderson, crowd work is a way to “automate” the job of a standup comedian, making the crowd provide your punchlines.) But crowd work clips, like ubiquitous man-on-the-street interviews and people menacing others in public for online clout, bother me primarily due to the power differential inherent in them, rather than the potential shoddiness of the work.

When you go on stage, as a standup, you are elevating yourself to a level of visibility. This is a choice you consciously make and how much (or how little) of yourself you actually make visible is part of that calculation. When you go to a comedy show, you aren’t necessarily expecting to be part of that show. And, if you are made part of the show, you aren’t planning on having that moment in time captured and packaged into marketing materials.

It doesn’t seem fair to me, the ways you are made visible against your will.

No standup I know enjoyed the crowd work clip boom of 2023. But, like man-on-the-street interviews, the algorithm loved them. If crowd work clips are the only ones being pushed, comics will create them. If man-on-the-street interviews are more likely to make you go viral, aspiring content creators will flock to them.

Arguably, the algorithm forced the visibility of audience members and random pedestrians, rather than any single individual.¹

Is it any surprise then, with the same people who crafted the algorithm making AI, it is used to force visibility on others?

Forced visibility should be part of AI’s calculation. But is very notably not. According to a report from UNICEF, 1 in 10 American teens know at least one friend or classmate who made non-consensual child sexual abuse material (CSAM) using generative AI tools.

No company is rushing to fix the situation they created, however. Grok, the AI integrated with Musk’s X, the Everything App, allowed people to generate 23,000 CSAM images over 11 days in January, according to Engadget. The breadth of abuse led Australian officials to say CSAM was “systemic” on the platform. Musk is now being sued by teen girls in the US and investigated by various European countries. But the only real change Musk made in the wake of these images being generated was banning Grok from posting the images on X. They can still be accessed and generated via the standalone Grok app.

AI imaging is primarily used to humiliate others against their will. Deepfake pornography, the “welfare queen” stereotype in full AI color, and the US government’s own AI ICE videos force individuals and marginalized groups into the fore to be humiliated and denigrated. These groups or individuals are forced into the cultural conversation, without any ability to push back against the false narratives created about them.

The algorithm and AI work in tandem to create a cultural milieu of adversarial relationships and humiliated participants. It seems to me the difference between human art and AI “content” must be the humanity, even if that moment of humanity won’t be allowed go viral on corporate owned platforms.

In making art in 2025, I am struck by the necessity for it to be in-person and participatory. But, the participatory nature of the art must also be cognizant of the power differentials between a performer and an audience. How can we create art that centers the humanity of all its participants? How can we create art that requests and invites the audience in, rather than forcing it? How can we create art that isn’t easily repackaged by forces outside our control?