Art, artifice, and the artificial

Can AI make something truly beautiful? The bar for makers is higher now. To risk not making a Mission Impossible 7 Pt. 2, but to risk being disliked. To make something beautiful, and true.

This weekend, I found myself rewatching a video that cycles back into internet collective consciousness every few years (or at least in the corner of the internet I live in—maybe this is revealing): the Goo Goo Dolls’ July 4, 2004 performance of “Iris” in Buffalo, New York, exactly 21 years ago. The setup feels almost mythic: rain veils the stage in thick, unrelenting sheets. Frontman John Reznik’s hair is plastered to his cheeks, black collared shirt soaked through. “Can you turn my guitar on?” He strums the intro as the crowd recognizes the song and screams. Then:

“And I’d give up forever to touch you.”

60,000 rain-soaked fans roar the words back.

Rain smears the camera lens in hexagonal blurs as if to say, “You had to be there.” The stage collects puddles. The band continues:

“When everything feels like the movies, yeah you bleed just to know you’re alive.”

The camera pans to the crowd, the real show, a mass of bodies, pressed together, also getting soaked by the rain. A few, seemingly ineffectual umbrellas punctuate the audience, but even they sway to the music. The rain is torrential, and seemingly beside the point. It is a scene at once dramatically cinematic and viscerally real. People look so freaking happy to be there, to feel the rawness of the rain in a shared baptism, a reminder that “alive” is a state but also a feeling.

By the final chorus, Reznik holds the mic out and lets the crowd take it home:

“I just want you to know who I am.”

There’s something so… earnest about this video that makes it powerfully nostalgic in today’s world of falsehoods and deepfakes. The whole performance is over-the-top, in-your-face emo, yet its irrefutable honesty transcends cringe: you cannot assail it. It feels like a glimpse of the “good old days” —as one commenter put it: “a performance straight out of a 90s movie”—not just because it’s two decades old, but perhaps because it reminds us of a time when surrendering to the moment felt simpler, easier, allowed, like an instinct instead of an effort. The crowd embodies the spirit of the lyrics, the willingness to get soaked to the bone so that the mess can be part of the memory. In that way, “Iris” is the perfect anthem for the grainy, rain-smeared footage being captured. The lyrics themselves were inspired by the film City of Angels, where an angel gives up his immortality to be with a mortal woman. Reznik has said he wrote the lyrics to reflect that desire for sacrifice: to give up everything (even eternity) for a chance to feel. As the song goes, “You’re the closest to Heaven that I’ll ever be.”

Watching this video makes me wonder about the promise of video itself, to deliver magic by giving permanent life to fleeting memory. But what happens when we seek more than just preservation and opt to manufacture memories instead? In our last issue, we talked about technology’s promise to help us rehearse for life’s big moments, to simulate them so thoroughly that uncertainty disappears. We asked: when technology can replicate the form of an experience (everything down to the lighting, the angles, even the faces), can the thing ever truly feel real?

My old boss had a favorite quote: “Sure, AI is cool. But have you tried skiing?” In other words, nothing feels like the thing itself when the thing is throwing your physical body down a 3,000 foot mountain. That’s the thing about the Buffalo video, too: it's compelling not because it's technically perfect, but because it's so clearly, and beautifully, lived. Indeed, the rain frequently blurs the image, a reminder that our experience is mediated by the camera. We watch not to experience the performance directly, but rather to imagine what that must have been like.

It’s worth asking what makes a moment feel beautiful in the first place. The famous Keats line goes “Beauty is truth, truth beauty,—that is all / Ye know on earth, and all ye need to know.” Like truth (because it is truth), beauty is impossible to define precisely because it resists being pinned down. It is not the tidy harmony in Bach’s fugues, but the risk of collapse at every measure, demanding the musician to hold two conflicting voices in fragile conversation. It isn’t quite the polished stillness in Vermeer’s interiors, but the single shaft of light revealing dust in the corners, the admission of humanity, of truth, that is beautiful.

Even in film, beauty often lives in the scenes that feel ineluctably real. I (finally) watched Anora this weekend. For two hours and nineteen minutes, I was swept into electric red and blue hues and addictive pacing that oscillated between frenzy and stillness. But it was the ending—the stark collision of light and dark (fitting, for a name that literally means “light”) as we sat in the creaky, dingy car with Anora—that at once devastated and consoled me. In the quiet of that final scene, the human condition shines beneath its thick patina. This is the truth of our existence; this is the cost of wanting anything at all.

In Reclaiming Art in the Age of Artifice, J.F. Martel reminds us of Baudelaire’s simple provocation: “Le beau est toujours bizarre.” The beautiful is always bizarre. It resists easy interpretation, unsettling you even as it moves you.

Martel goes on to warn us that most of what we call “art” today is really artifice. He discriminates between true art and two forms of false art: the didactic, which tells you what to think (the anxiety medication commercial with a woman prancing through dandelion fields while disclaimers scroll by), and the pornographic, which gives you exactly what you want to see (a video of Sydney Sweeney lounging in a bathtub selling Dr. Squatch Sydney Sweeney Bathwater Bars). “The last thing artificers want,” he writes, “is to divide their audience. Their competence as creators hinges on their ability to replicate the same emotional response in as broad a demographic as possible.”

It’s easy to see where I’m going with this—technology easily enables the flattening of our tastes, making beauty feel cheap and everywhere all at once. The layup answer is that all output of technology is artifice. Facetune smooths our blemishes and nose bridges into a template of algorithmic attraction. Manufactured content, like mass-produced Wonderbread, but now with our faces. The real danger is kitsch, which Milan Kundera so memorably captures:

“Kitsch causes two tears to flow in quick succession. The first tear says: How nice to see children running on the grass!

The second tear says: How nice to be moved, together with all mankind, by children running on the grass!

It is the second tear that makes kitsch kitsch.”

Kitsch is not life itself, but life repackaged, scrubbed of shit and contradiction. As Kundera sardonically notes, “The brotherhood of man on earth will be possible on a basis of kitsch.”

I’ve thought about these problems while working this summer in AI-generated video. At Hedra, we’re building tools that let you create entire performances from a single image and line of text. It’s astonishing technology, capable of both art and artifice. I don’t have to look far for cautionary tales. A brief walk down memory lane takes us to the recent wave of AI-generated anime clips attempting to mimic the style of Studio Ghibli. The videos generated were slick and pretty, even awe-inducing the first time, and it was easy to grow usership on the coattails of the trend’s virality. But when Hayao Miyazaki himself called AI-generated animation “an insult to life itself,” he wasn’t talking about fidelity to form, but the lack of creativity and personal stake involved in AI mimicry.

At the same time, indie filmmakers say they can now create scenes that they never could have afforded before. These use cases include the practical—like being able to film at an airport terminal for a fraction of a dollar through a SaaS subscription versus a six-figure contract—and the surreal—like visualizing novel magic systems and introducing time as a character itself, unfolding and folding in on itself. Generative AI has the potential to democratize the deep-pocketed language of cinema and place bigger visions in the hands of smaller teams.

But these tools are also only as good as the stories we choose to tell with them. Why does the character glance off screen, tug gently at her pocket, pause mid-sentence, then divert her glance at the wrong moment? Is there consciousness behind her gestures? Or is “she” but a series of ultra-realistic movements with no emotional context, trapped in the uncanny valley?

My current (biased) answer is not to reject these tools. It is to ask much more of ourselves as makers. To risk not making a Mission Impossible 7 Part 2, to have the courage to be ugly and disliked, to risk being the belle of only a single beholder’s ball. The bar is now higher to experiment with what is available, and to familiarize ourselves with the levers for creativity, so that we might make something that feels—is—real. Something beautiful and true. I want to feel unsettled, divided, even, but most of all moved.

In the end, the point is the same as it’s always been:

“When everything feels like the movies…you bleed just to know you’re alive.”