Smack dab in the center of this year’s SXSW festival in Austin, Texas, there was an enormous grime gap in the floor, blocks large, the place there was a conference middle. The pageant’s occasions continued round it in inns, however the constructing’s absence was like a lurking image. Of chaos, of disruption. Of the world in 2026, coping with AI and the whole lot else.
I don’t know what the remainder of 2026 will carry, however the vibe I felt at a vibe-filled present made me query how AI can work with our lives, our artwork and our existence. Instead of preventing it, the convention awkwardly embraced it and challenged it. I noticed pockets of labor throughout the place and questioned about it. Conversations. And the right way to escape it.
Everyone’s attempting to deal with a world that is all of the sudden method too overloaded with AI, producing paperwork, photos, deepfakes and music, injecting assistant brokers into our working programs, even launching total unleashed and interconnected agent systems all speaking to one another on their own social networks. Job-threatening, consistently shifting, coaching on our information and aiming for our faces. Do we run from it, attempt to destroy it, or use artwork to query and problem it?
SXSW gave me numerous the latter, in several slices.
In my panel at SXSW, Meow Wolf’s Vince Kadlubek and Niantic Spatial’s Dennis Hwang defined their experiments overlaying tech into artwork in bodily installations. Kadlubek discussed how AI’s infinite slop artistic device turns into uninteresting over time, whereas intentional artwork counteracts that. And that is precisely how I felt transferring by deliberately made experiences that turned my ideas about AI inside out, all in several methods.
Fabula Rasa: Dead Man Talking is a VR recreation with characters that use generative AI to talk by way of Claude. It’s quite a bit funnier and extra profitable than I imagined it could be.
AI seeping into our gaming chats, for higher and worse
In a VR headset in a lodge ballroom, I chatted with cartoon fantasy characters in a whimsical recreation referred to as Fabula Rasa: Dead Man Talking, made by recreation studio Arvore. I may make any request or beg as a lot as I wished from my cage, the place I was held prisoner for offending the King and stored dangling over a monster’s mouth for execution. Could I plead my case to them? The cartoonish VR characters responded, however by way of generative AI improvising off a script from a writing group, utilizing Claude.
The chats have been enjoyable, ridiculous. I made myself an irresponsible magician and leaned into improv with the characters who approached me. None of them dissatisfied, which is a shock for dialogue that is considerably AI-generated. Most interactions felt frazzled and absurd, however it labored for the model and the humor of all of it. There was a little bit of a delay for responses to kick in, although, standard-issue for lots of AI conversations.
This was the greatest use of AI I noticed. But what may it imply for future video games, like RPGs? It’s an unsettling thought in the event you’re a author…or, thrilling. Indie video games may find yourself discovering methods to department out responsive dialogue in ways in which nonetheless really feel custom-written and crafted. I do not know.
On the much less profitable finish was Love Bird, an interactive recreation present expertise directed by Cameron Kostopoulos. I was wowed by the preliminary onboarding, the place the “producers” referred to as me on my telephone to interview me. The producer was truly an AI chatbot with a surprisingly speedy response time. I satisfied the AI to be a participant, after which was led right into a room the place I spoke by way of Xbox controller and headset microphone with a PC recreation on a monitor, the place I was competing with others whereas carnivorous bird-people threatened to eat us. I’m unsure why, precisely. And I do not know the way it all ended, as a result of my chats with the host and contributors fell into damaged loops that made us must give up out early.
Love Bird was fast-paced and responsive, but additionally too chaotic and bizarre, even for somebody like me who likes bizarre. It did not really feel prefer it was actually being attentive to me, and I did not really feel like I had house to course of. Maybe that is by chaotic design, however after rising, it simply made me wish to really feel much less AI-spammed and have video games that did not flood me with as a lot dialog as this one did. I wanted a quiet house. My favourite immersive experiences are sometimes the quiet ones, not the chatty ones.
Me and my AI portrait completed in the model of Jonathan Yeo, at an AR glasses-enabled exhibition referred to as Spectacular.
AI as a private transformational lens
In one room, I stood at a podium and skim a portion of New York Mayor Zohran Mamdani’s acceptance speech from November as, earlier than me, video clips of crowds cheering performed on a big video monitor, seemingly reacting to me. A couple of minutes later, I heard my voice delivering extra of Mamdani’s speech, AI-generated in my voice, to movie clips of inspirational moments of assist. I noticed my very own face layered into the background of a few of these clips, too.
The Great Dictator, directed by Gabo Arora, is a museum-style participatory exploration of the energy of rhetoric, provocatively named for the Charlie Chaplin satire about Adolf Hitler. The three speeches you’ll be able to select from — Mamdani’s, President Ronald Reagan’s on taking down the Berlin Wall, and Malcolm X’s The Ballot or the Bullet speech — are all picked to signify highly effective moments in historical past, and the exhibit is about embodying historical past and feeling the energy of speech and rhetoric in a private method — and regarding it from a brand new, private, and possibly extra empathetic angle. The voice AI was generated by ElevenLabs, and the video clips at the finish have been hand edited, however with AI overlays of my face dealt with by Runway. What shocked me was how a lot I ended up being in historic paperwork. Is this a deepfake? Is it embodiment? Is it each?
The podium I spoke at for The Great Dictator, an expertise at SXSW that overlaid my voice and face into historic paperwork.
Another artwork expertise embedded me into the work: Spectacular, by Jonathan Yeo. Yeo is an artist from London whose portrait work consists of King Charles III, President George W. Bush and designer Jony Ive of Apple renown and has performed with tech in a lot of his installations. This gallery at SXSW, replicated from an exhibit that was in Paris earlier than, used Snap Spectacles AR glasses to soften the actual portraits with augmented results and voice narration from Yeo. And, in a while, the portraits started overlaying my very own face, remodeled in artwork types that matched Yeo’s utilizing generative AI educated on his work. At the finish, I acquired a printout of my portrait, “signed” by Yeo himself.
I spoke with Yeo in Austin after experiencing his work. He admitted that AI is a provocation right here, however that he needs to personal the course of that AI is attempting to take from our personal information in every single place. And he is attempting to use AI and AR in ways in which really feel intentional and refined as methods to assist play with and produce the artwork to life, in museums and elsewhere. But once more, like with The Great Dictator, I questioned: How a lot will “permanent” paperwork of artwork and historical past start to soften over time with AI? What will likely be stored intact, and who will implement the line?
A desk of things I had assembled in the AI-guided expertise Body Proxy, and the video of me (an AI-generated pseudo-me) celebrating.
AI as damaged manipulator
Wearing a pair of Meta Oakley good glasses, I stood in a room filled with objects on cabinets as a voice directed me to open a drawer, discover a greenback invoice there and put it in a shredder stuffed with invoice fragments. I did it. The AI remarked with nice shock at how compliant I was. From there, I “competed” duties to show my worth as human labor, graded by an AI that noticed my actions by the glasses digicam and confirmed my stats on a TV display screen, together with a deepfaked dancing model of myself.
Body Proxy, by Tender Claws, applies Meta’s glasses digicam feed into its personal artwork AI app on a telephone to discover how AI may make us proxies for bodily labor. It’s bizarre and satirical like a few of their different VR work (the recreation Virtual Virtual Reality, amongst others), but additionally pushes at a a lot larger query: How a lot is AI breaking us or manipulating us? How a lot are we prepared to be manipulated?
Escape The Internet (Part One), an interactive recreation I performed in a movie show at the Alamo Drafthouse, turned comparable concepts of manipulation right into a social experiment. Created by Lucas Rizzotto, one other VR/AR provocateur artist, it concerned no headsets or glasses. Instead, everybody in the theater used their very own telephones to hook up with a personal server that “ran” the recreation and gave us little private avatars, feeding us surveys to gather our private tendencies after which having us play social voting video games to see how we would polarize on selections like, as an example, who to kill: one one that shared our political beliefs, or 5 who did not?
Holding up my telephone as I linked my system to Escape The Internet’s group recreation in an Alamo Drafthouse movie show.
It’s all absurd and humorous and guided by Rizzotto’s in-person steerage at the entrance of the theater, and alongside the method, I thought of how social platforms manipulate us with algorithms. Here, on this room collectively, we’re inspired to seek out one another, acknowledge one another and love one another. The expertise has branching paths and might be replayed, and will re-emerge in future conferences and occasions. But, once more, I requested myself: How a lot of AI is a recreation that is taking part in me, as an alternative of me taking part in it?
A fuzzy, purring robotic idea referred to as Kabbage, which I held at the robotic design discuss I attended at SXSW. What may a reference to a robotic (or AI) really feel like?
Design for AI remains to be unfinished (or nonexistent)
In a few of the panels I sat in on, and in conversations I had, I acquired a creeping sense that AI is transferring too quick for artists or ethicists — or anybody else, actually — to cease and correctly course of. One panel exploring The Future Design Language of Robots, with Olivia Vagelos of the Design for Feelings Studio, and Savannah Kunovsky, managing director of Ideo’s rising expertise division, tapped into the assumptions we make about robots. I teamed up with somebody subsequent to me to attempt to dream up concepts to interrupt my assumptions and assume freshly about what robots could possibly be.
Kunovsky and Vagelos each agreed that designing for AI presents comparable challenges proper now, significantly as a result of the tech is transferring too quick for design to correctly attend to it. But sadly, my try and report what they mentioned as a quote was sabotaged by my AI-enabled Meta Ray-Ban glasses, which activated as the microphone when I tried recording a voice memo from the panel on my telephone, muting the audio fully due to noise cancellation. Wearables are nonetheless damaged, too.
Another panel, referred to as Generative Ghosts: AI Afterlives and the Future of Memory, led partially by two Google DeepMind researchers, mentioned many desirable angles on how we will responsibly deal with archiving our lives by way of AI as reminiscences in the future, and who controls that capability. The panel had no particular solutions however loads of questions. And, as my very own try at recording it was additionally erased by my activated good glasses, it gave me an extra stage of absurd friction which made me surprise: Will these archived reminiscences ultimately be misplaced, too, from large tech firms that sundown companies or introduce noncompatible codecs, memory-holing the reminiscences?
Making a collage with pals that used no AI or computer systems at all.
AI is threatening, however typically not profitable in fulfilling its guarantees (or threats). Self-driving Waymo automobiles flooded Austin throughout SXSW, with my Uber app typically pushing them on me as an alternative of human drivers. I gave in and took a couple of for amusement, however they normally took longer to get the place I was going. And, one unlucky night, my Waymo took a bizarre roundabout route that ended up dropping me off a half mile from my vacation spot on the fallacious facet of the freeway.
My favourite SXSW reminiscence was making an old style collage out of journal clippings with pals at an artwork gallery over wine, one thing that concerned no tech at all. We labored our magic with instinct, scissors, outdated magazines and good dialog. Was it excellent? No. But it value quite a bit lower than generative AI. Which additionally makes me surprise if all these AI instruments being supplied to reinforce or supplant creativity are needed, or whether or not we’ll simply rediscover that we had extra instruments than we realized all alongside.
