AI Killed Images. Legacy Russell Knows How We Can Revive Them

AI Killed Images. Legacy Russell Knows How We Can Revive Them 1

Within the frame of a single image lives a lifetime of experience, emotion, and affect. Photographs trap drama and performance. TikTok videos delight us with humor and consequence. For Legacy Russell, author of the new book Black Meme: A History of the Images That Make Us, virality may be the best indicator into how an image speaks. In deciphering the speed and movement of a given image, the way it courses through different contexts and communal spaces, its architecture of being becomes all the more manifest.

Russell tells me that “asking questions about the truth of images, consent, and how we are complicit in the way they move—these are all things that are critical to this core driver of representation, and what the implications might be.”

Those implications, of course, are all around us and only growing in intensity with the spread of generative AI (during a crucial election year, no less). Already we are witnessing a “decay of imagery,” she says, a degradation of what is accurate and authentic.

To learn more, I phoned Russell, a New York City–based curator and author (her first book, Glitch Feminism: A Manifesto, was published in 2020), hoping to better grasp how images move through our world today, particularly across the internet, as they assume novel digital forms and sometimes strange meanings—or a lack thereof.

Jason Parham: In the book, you call the Black meme both trap and trope—why?

Legacy Russell: The frame of memetic culture or viral culture is oftentimes seen as its own economy. Often within that, there is the expectation that if something goes viral, or if meme material travels from place to place, there is something within that that is a positive. It is about a kind of recognition of a value set and a material having meaning to a broader public. This is why the book centers on moments that are not specifically tied to the internet, but predate the internet as we know it now.

Yes—predate and predict.

The trap is that in the acceleration of images there actually is not an assumed economy that allows for folks to engage questions of authorship or provenance in a way that has a robust infrastructure to support Black thought and Black cultural production. In fact, many things that are traveling through viral material can be turned on their ear within a model of dispossession, economically or otherwise. This is where trap and trope have their own relationship.

Speaking to those pre-internet moments, you suggest that Black memetic culture would not exist without the gruesome history of lynching photography. How did those spectacles inform our relationship to memes today?

When you asked, what’s an early meme I remember, for me coming into awareness of lynching postcards was that nascent memory. The very idea that this was something that could exist as a material. They exist now as archival material but also as a fungible economy; these are items that are considered American collectibles. Folks buy them online. But this is how we better understand the infrastructure of images and text.

In what way?

If memes are intended to circulate across boundaries and through different publics and intimacies, lynching postcards were among the first memes. Lynching postcards traveled across state lines. They literally were a form of terrorism in how they allowed there to be documentation of a violent act that was sanctioned by a white public. And celebrated. As well as preserved as a souvenir.

Think of a collector who is writing a postcard to a loved one from a position of endearment, and on the other side of this correspondence, they are sharing an image of someone being brutally murdered.

We should be asking about similar models of engagement and complicity that are enacted in this moment now when we are looking at things via our screen, exchanging materials of violence against Black people. The question, really, is how that is sustained by a broader public as part of the economy of the internet.

Because of the rate at which Black images move on the internet, so swiftly robbed of all context, is there a way to reclaim ownership over them?

Part of the interest in writing Black Meme, and having it exist as a physical [book] away from our screens, was the idea of creating slow media. So the slow media that has then intersected with the accelerated media of now—cyberspace and digital space—is very critical because part of this discussion is about a mitigation of speed.

What happens when things are circulated and compressed in a mass acceleration, when many things are being perceived as copies of copies? Images are degrading over time. Often we are not seeing them for what they are. We’re seeing them as a contour instead of actually better understanding and situating the truth that lives inside that transmission. What I am proposing is a need to engage a different model of thought around the transmission of Black culture, as well as Black people and their representation through and beyond our screens.

The inverse of slow media is our current reality. We are bombarded with media at all times. The velocity is constant and unpredictable. What is the danger in how social media, specifically digital spaces like Instagram or TikTok, has shaped our understanding of how memes live or die?

It’s important to ask questions about the sustainability of the Black meme.

Sustainability, yes.

Vine in particular was a place where a lot of Black people were creating space and community, as well as engaging performative action. I call it performative action because it came in different forms of sonic engagement, movement, and models of gesture. The very idea of Vine was essentially a site where Black people were putting this to good use, and then it reached a point where it couldn’t sustain itself any longer, and thereafter was the rise of Instagram. Now TikTok is part of that broader equation, that brief and furious history.

Platforms like TikTok are increasingly under scrutiny because they are being used as sites of organizing and exchange for and by people of color across many different diasporas. As the economies shift around that space, certain platforms fail to exist or collapse altogether, the question becomes, who is that impacting?

Generative AI seeks to automate every part of our lives. In Black Meme, you call for royalties and reparations around those issues. What does this mean for the images that we will increasingly encounter, specifically within the Black visual medium? Are they under threat or is a new window opening?

That’s a big question. It might be useful to complicate rather than simplify.


As we’ve seen with early histories of digital automation, there have been many think pieces written around the feminization of automated labor, and what that looks like. Be that the feminized frame of Alexa or Siri. Aspects of these technologies exist within the gender and class economy, and also the race economy, although that’s spoken about less frequently. There has always been an imprint that is gendered inside of technologies as a means of making them familiar to us.

How do you see that coming to the fore with the current uses of AI?

When we ask what does it mean when there are AI resurrections or imitations of popular songs or performances across different media or artworks, part of the risk really is about the replacement of certain models of labor and what impact that has on a broader economy in the world.

Right. AI can mimic almost any famous musician or artist now.

It also encourages us to think differently about digital Blackface. AI is allowing for a different chapter of that to unfold at an accelerated rate. It also allows for people to consider differently about how they present different types of truths and/or fictions in the performance of race as it is resurrected in the world.

Which can get ugly real quick, as we’ve already seen with so many grotesque depictions online.

This is core to what you’re asking. Some of the defense around cultural production in relation to Black people has always been that there are people behind that work. We think about the economy as reparations or royalties and ways in which aspects of these things can be mitigated.

At this moment now, there is a wide proliferation of people testing out different approaches to labor overall. Think of the Black people who are inside of AI in advertising, in reportage, even in thinking about a broader cultural discourse and visual representation of Black people in general stock imagery.


So when we think about what is sellable, there are calls for an equitable form of representation that accurately intersects with an ongoing blending of the world, where more and more people of the diaspora are situated at the center of how media moves. While at the same time, some of the core forces in the economy that are rising up to extract from that are very aware that the representation of Black people can be done and engineered by AI.

In the years ahead, if equity is too much of a risk, or is too expensive, Black people will be taken out of the equation altogether.

We are already seeing that.

What does it mean if AI does that work? Which essentially is to create ideas of human beings that do not exist but that can sell products, and can advance certain ideological agendas, but can also perform labor without breaking a sweat, needing a break, or requiring compensation, and in fact having that be something that can create economies for non-Black creators in ways that are perhaps even unknown to us in this very moment. That right now is the conversation.

Which will only get more tangled as AI advances.

The discourse around reparations will complicate things further because it will be not purely about how the human has been situated in the middle and dispossessed, but rather how the human is set apart, marginalized, and cut out of the equation altogether with a different model that can form this work. And what does the compensation framework look like there when in fact it is, as I noted prior, a kind of copy of a copy? When Blackness itself becomes its own performance-based material and that itself is its own engine.

Did you see the “All Eyes on Rafah” AI image that went viral on Instagram?

I did. The use of AI to bend reality or to narrate outward fictions altogether can be corrosive to our understanding of what is or is not true or real. How this is applied can be a glitch or strategy in a broader system—in this particular case, using an AI image to subvert algorithmic censorship to elevate a discourse around ongoing violence that requires global attention is a double-edged sword. While on the one hand it may broaden a consciousness, on the other it aestheticizes a struggle for liberation and makes for quippy viral material to be exchanged.

I do think some people had genuine intentions, but it mostly read as performative empathy.

The image itself establishes a reality that lives far from the reality of what is happening on the ground.

A reality many people blindly and happily live in.

Its use is complicated. The AI image conceals the mechanics of violence and oppression as they are occurring in real time while at the same time seeking to advance awareness of a human rights emergency.

This is an example of how material can be used toward and against its own mission. An example of how images, as they live “inside” of other images, can be part of the strategy of shaping ineffective attention economies, which fuel fantasies of hegemony but don’t always honor the real lives of the people living “inside” the idea behind the image itself.