AI pop culture is already here

AI Video & Visuals

Last month, a YouTube user named demonflyingfox uploaded a video titled “Harry Potter by Balenciaga.” Harry his Potter movie characters (Hagrid, Ron, Hermione, Snape, McGonagall, Dobby) as lanky models with aggressive cheekbones (somewhat cheesy) in gothic cloaks and leather jackets was drawn. Catwalk-worthy electronica To his beat, the actors blink, nod, and speak lines from books remixed with fashion references. “You’re Balenciaga, Harry,” Hagrid says, instead of delivering the news that Harry is a wizard. In 3 weeks, he almost got 5 million views. Released less than a week later, the sequel has grossed over 1.5 million. His pop culture mashups of one famous thing with another are the archetypes of internet meme-making. What makes “Harry Potter by Balenciaga” unique is that it was generated with artificial intelligence tools. According to the creator of the video, Alexander Niklass, a Berlin-based photographer who created the demonflyingfox channel, the video is a new way for his AI to “create cinematic moments.” It shows ability.

AI tools were involved in each step of Niklass’ process, as well as each element of the video. He created rudimentary still images in his Midjourney, reminiscent of Harry Potter actors and costumes through his prompts with text such as “male model, grotesque, Balenciaga commercial.” He then used his ElevenLabs (a “voice duplication” tool) to create models of the actors’ voices based on previously recorded voices. Finally, he submitted the image to his service called D-ID. This service is used to create “avatar videos”. The service is subtly animated portraits that aren’t all that different from those that appear in the Potter World newspapers. D-ID added a signature lip his sink and head nod. Niklas explained that it was a reference to a fashion model with her chin tilted towards the camera. The combination of children’s movies and adult high fashion had no particular symbolism or artistic intent. It’s “entertainment,” Niklas said. But the most compelling aspect of this video may be its emptiness, a nonsensical clash of cultural symbols. Nonsense is the point.

An AI tool could have replicated an actor’s face and generated fashionable outfits, but only Niklas could come up with this concept. It required a keen observation of both high fashion and the wizarding world, as well as a very specific and very online sensibility. humor. With tools like Midjourney open to everyone online, “everybody can now create something visually appealing,” he says. “But the AI ​​still can’t generate flavors,” he continued. “Taste” means “good aesthetic judgment.” That means background knowledge of what you’re producing and a sense of what looks good without falling too far into the uncanny valley. In other words, execution may have been democratized by generative AI, but ideas have not. Humans are still the creators, editors, and curators of AI effects. Evidence of Niklass’ taste can be seen in the many imitations of his videos currently on YouTube. Anyone can access the same technology and try replicas according to the formula he sets. (A video tutorial made by an AI education channel called PromptJungle shows the exact process.) There are videos for “Matrix by Gucci,” “Star Wars by Balenciaga,” and “The Office by Balenciaga,” but which one? is also not very attractive. Strange as the original.

No one would mistake ‘Balenciaga’s Harry Potter’ for real-life footage, but other than perhaps the actual fashion ad campaign, another AI-generated image recently made the headlines, which is very disappointing. Because many people thought it was real. It was, or looked like, a picture of Pope Francis walking down the street wearing Baroque her version of her jacket, a silky white puffer with a high collar and hood. A crucifix necklace hung around her neck and she was carrying coffee to take home. “Swagout” is an apt description. The image detail and texture made it look very realistic. When I first encountered it, scrolling quickly, I thought it was real and barely noticeable. Of course, the Pope wears fancy clothes.Chrissy Teigen, model and actor murmured “I didn’t give it a second thought,” she said. But as he told Chicago, it was created using mid-journey by a man from Chicago who called himself Pablo Xavier when he was there. tribune, a lot of mushrooms. He said, “Pope Francis. Balenciaga puffy coat. Streets of Paris.” .)

Images, created unintentionally other than by the whim of the imagination, were very powerful because they were made to look like photographs. negates the effort involved. (A real-life version of the “Harry Potter by Balenciaga” video could have cost millions of dollars in talent alone.) We pondered this discrepancy between input and output and interacted with ChatGPT. It was more Socratic than an actual discussion, more like talking into a mirror. I asked how they changed it. It replied that “the line between the real and the artificial is blurring”. created it, so isn’t it true that it’s “real”? This idea of ​​’fantastic realism’ seemed to fit me perfectly. AI content has the appearance of realism with only stylistic reality, not actual reality.

In a 2022 interview, Midjourney founder David Holz used the phrase “aesthetic accelerationism” to describe the massive amount of generative imagery made possible by public AI tools. It evokes a world where every style, every idea, every possible remix is ​​generated as quickly and frictionlessly as possible. Like ‘Balenciaga’s Harry Potter’ or the daring pope, the successful ones stick around and grab attention. It may be less because they are artistically brilliant than because they have worked out the formula of attention. are seamlessly integrated. Or, like a trompe l’oeil, we can easily look at them even though we know they’re not real, or we’re driven by the fascination of perfect replicas that are easily recognizable.

Over the past few days, I’ve been looping a new hip-hop song over and over. Called “Savages” by the French outfit AllttA. The song is sweet and nostalgic, featuring synthesized strings and a snare backbeat. Jay-Z features in his Throwback his style what sounds like a trade poem with his AllttA’s Mr. J. Medeiros. Of course it’s not Jay-Z. His AI model of his voice, presumably used without the artist’s permission. This is another example of fantastic realism. The human-written songs are good in their own right, and they’re perfectly fine without the fake Jay-Z, but a familiar voice adds something unspeakably compelling to this track, Sounds like his unreleased B-side from the 1990s. He has over 200,000 views on YouTube. “The idea of ​​enjoying this and that AI of his is beyond me,” one user wrote in a comment. I feel a similar existential turmoil. It sticks in your mind like an unsolvable puzzle. The AI ​​quality is good enough that a non-expert like me can hardly tell the difference. But the Rubicons seem to be crossed. It doesn’t matter if these artifacts are AI-generated or not. You can enjoy it as it is. It happened sooner than I thought, but now that AI-generated pop culture has gone mainstream, it seems unlikely to get rid of it.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *