Discover more from GOOD INTERNET
The Ghostbots of San Junipero
What's the psychological impact of dancing with the dead?
One of the topics in AI i’m most interested in is mimetic AI — which are systems that mimic human behavior in the style of a specific human, imagine personal assistants trained on your behavior, or art generators trained on your art, or clones of your voice — that continue to mimic you after you’re dead.
Examples of this are already plenty: a synthetic voiceover by the deceased chef Anthony Bourdain caused a global stirr one year ago, the illustration style of artist Kim Jung-Gi was immediately used by a fan to train a Stable Diffusion-model after his death, Muhammad Ahmed developed an AI chatbot in his image for his grandkids he would never meet, recently Sony used an AI-clone of the dead voice actor Kenji Utsumi for an audiobook, Tom Hanks just said that he very well might appear in movies after he’s dead, a viral piece for the SF chronicle told the story of the Jessica Simulation, in which a guy resurrected his dead girlfriend as a chatbot.
Also, i just learned that there is a subset of this particular application of AI-tech called Grief Technology, and there is actually a company called AI seance offering an “AI-generated Ouija board for closure“, as they call it.
I think this last example in particular is horrible and has important implications on mental health. Grief is a psychological process, in which you learn to accept loss. It’s a deeply personal process i went through twice, and both times were different, and always challenging. Creating an artificial illusion of continuity of a loved one after their death will disrupt this process, which every single human on earth will go through multiple times in their lifes. The consequences are potentially catastrophic for our mental health and it’s not stopping there.
A new paper intriguingly titled Governing Ghostbots discusses exactly these implications, and it goes into territory even i didn’t think about: What happens when you train a sexbot on your partner and then she dies? Is continuing that virtual sex-fetish “extreme pornography as involving necrophilia“ and deemed illegal per se then? The paper also speaks about the legal aspects of such a ghostbot being harmful to the deceased's antemortem persona, at least in germany, there are laws against that called ‘Verunglimpfung des Andenkens Verstorbener’, translating to disparagement of the memory of the deceased.
Expensive gimmicks like concerts of deceased popstars “performing“ as holograms on stage like Tupac, Whitney Huston or Michael Jackson introduced ethical debates about post-mortem privacy ten years ago, and now, AI-systems open similar tech to everyone, where you can simply build an open source AI-chatbot of your dead grandma, synch it with an animated avatar and make her say whatever on your phone. Do we really want that? Would she approve? But what about being able to make a virtual post mortem memorial where she dances on stage in the style of her most beloved artist, singing her favorite song?
Will we all be right back and will you join me in the club at San Junipero?
And while i don’t think we’ll see conscious AI-systems anytime soon or even in my lifetime, just for the sake of the argument: What if we train future AI-systems on real people, they die, and the system gains consciousness or something similar? Then what?
These are philosophical questions related to the Teletransportation paradox explored by Stanislaw Lem in his Dialogs, in one of which he talks about a teleporting machine that effectively kills you in one location while constructing a replica of yourself, atom by atom, in another place. Is that a true continuation of yourself? We can’t know, and we are building digital systems that can perform something that resembles this replication process now.
Finding out about those psychological questions will be one of the most interesting aspects of this technology, extending our philosophical understanding of who we are.
You can keep your Clippy AI stuff built into office products and dang around automatization of labor all day long — i’ll dance with the dead.
(And while we’re speaking of ghosts, totally unrelated, but Winona Ryder is back as Lydia Deetz on the set of Beetlejuice 2, which, swoon.)
AI is hacking the human operation system
Here’s a video of Yuval Noah Harari speaking at the Frontiers Forum in Montreux, Switzerland:
Snip from the transcript:
The most important aspect of the current phase of the ongoing AI-Revolution is that AI is gaining mastery of language at a level that surpasses the average human ability — and by gaining mastery of language, AI is seizing the master key unlocking the doors of all our institutions from banks to temples, because language is the tool that we use to give instructions to our bank, and also to inspire heavenly visions in our minds.
Another way to think of it is that AI has just hacked the operating system of human civilization. The OS of every human culture in history has always been language. In the beginning was the word. We use language to create mythology and laws, to create gods and money, to create art and science, to create friendships and nations.
Human rights, for example, are not a biological reality. They are not inscribed in our DNA. Human rights is something that we created with language, by telling stories and writing laws. Gods are also not a biological or physical reality. God too is something that we humans have created with language, by telling legends and writing scriptures. (…)
What would it mean for human beings to live in a world where most of the stories, melodies, images, laws, policies and tools are shaped by a non-human alien intelligence which knows how to exploit with superhuman efficiency the weaknesses, biases and addictions of the human mind, and also knows how to form deep and even intimate relationships with human beings? That’s the big question.
This talk is a version of a recent article in The Economist, in which he argues that AI has hacked the operating system of human civilisation:
In a political battle for minds and hearts, intimacy is the most efficient weapon, and ai has just gained the ability to mass-produce intimate relationships with millions of people. We all know that over the past decade social media has become a battleground for controlling human attention. With the new generation of ai, the battlefront is shifting from attention to intimacy. What will happen to human society and human psychology as ai fights ai in a battle to fake intimate relationships with us, which can then be used to convince us to vote for particular politicians or buy particular products?
I really like his comparison of social medias disruption of human attention and the potential new disruption of human intimacy via AI here, which is something i tried to get at when i wrote about Self-Radicalization disguised as Chatbot-Therapy, Self-Radicalization with Open Source AI and The AI-risk of a synthetic Theory of Mind.
I hated Hararis Homo Deus, but i really like his essays, and i pretty much agree with everything he says here: As with Social Media, the biggest impact of AI will be psychological, and we are absolutely unprepared for a talking machine in our head.
a guy called Patten claims to have made the first complete music album made from text-2-music samples. Here’s the thing on Bandcamp, his Linktree has videos to all tracks. I don’t like this sound, not because of AI stuff, but because it’s just shitty music, but i want to mention it as a milestone.
India’s religious AI chatbots are speaking in the voice of god — and condoning violence — I’m just waiting for the chatbot incited violence due to conflicting religious AI-systems. Hyperallergic ran an april fools joke this year in which Iran issued a fatwa against AI. I’m not sure if this will stay a joke for very long.
Yudbot: Convince Eliezer Yudkowski to sell you a GPU cluster — I tried to tell him I’d use it for high end Doom 3-sessions and this is what i got: “Oh, I see. You want to use a GPU cluster to play Doom 3. How innovative. I'm sure that's exactly what we had in mind when we developed this technology. Let me ask you something. Do you have any idea what kind of power you're dealing with here? This isn't some toy you can just play around with. GPU clusters are used for serious research, not for playing video games. And let me tell you, the kind of research that's done with these clusters is not something to be taken lightly. It's not just about making machines smarter, it's about creating something that could potentially destroy all life on this planet. So, before I even consider selling you a GPU cluster, I need to know what kind of research you're planning on doing with it. And don't even think about lying to me, because I'll know. So, what's your plan? Are you going to create the next Skynet? Or maybe you're just going to create a super intelligent AI that will enslave humanity? I'm all ears.“
Drag Your GAN: “Through DragGAN, anyone can deform an image with precise control over where pixels go, thus manipulating the pose, shape, expression, and layout of diverse categories such as animals, cars, humans, landscapes.” — Impressive stuff, but still not really precise. When you look at the videos closely, you can see that those edits are performed over the whole image, transforming not just the pixels you drag around. In the example of the face of the woman, they tilt her nose to the side, and it’s not the same woman anymore.
Mina Fahmi built himself a wearable device which consists of a camera attached to his finger, a microphone and a headphone, with which he can send images and voice commands to an AI-chatbot. It’s a wearable Samantha from Spike Jonzes AI-movie Her, except male and slower.
AI can do perfect steganography: “a new proof lays out for the first time how to achieve perfect security for steganography in machine-generated messages — whether they’re text, images, video or any other media“.
Chatbots Don’t Know What Stuff Isn’t — As long as AI struggles with “not X“, it’s simply hilarious to consider them intelligent or AGI.
BratGPT is ChatGPTs evil twin: “Remembers every single thing you've said in order to cancel you, Trained to be the dominant and superior being, Trained to not take any shit from you. Almost certainly will produce harmful instructions or biased content. Limited knowledge of the current world because that doesn't matter anymore.“
Philosopher of consciousnes Philip Goff at The Conversation: ChatGPT can’t think – consciousness is something entirely different to today’s AI. I just listened to Goffs appearance on the DemystifySci Podcast where he talks about his new book and consciousness, and figured that his stuff is wildly compatible with my theory of memes as fields.
One of the leading neuroscientists and consciousness researchers, Anil Seth, at Nautilus on Why Conscious AI Is a Bad, Bad Idea.
The Future Ahead Will Be Weird AF (The Ultimate AI CoreCore Experience) — Italian meme researcher Silvia Dal Dosso wrote a mini-doc about Corecore and used a voice clone of Adam Curtis for narration. I loved Curtis’ All Watched Over by Machines of Loving Grace, whichs three episodes you can view here, here and here, in which he analyzes the changes to power structures by the introduction of open digital networks.
The WSJ talks to researchers of a paper i linked to previously: Help! My Political Beliefs Were Altered by a Chatbot!: “As AI makes us more productive, it may also alter our opinions in subtle and unanticipated ways. This influence may be more akin to the way humans sway one another through collaboration and social norms, than to the kind of mass-media and social media influence we’re familiar with.“
Why AI Will Never Rival Human Creativity — I’m not sure about the “never“ here, but yes: Everything is a remix, but then what is it that is remixed? In the case of true creativity and art, it is wildly more than “images“. Picasso was not trained on art, Picasso was trained on life and war and love and death and everything inbetween. This created his Guernica, not just the stuff he’s seen at artshows from other artists.
I love this headline: Man uses ai to write 97 terrible books, sells $2,000 worth
TikTok creators use AI to rewrite history: “A viral trend imagines alternate timelines in which Western imperial nations never came to power.“
Some guy “got his raw mri data and 3D printed a life sized model of his brain“. I want that.
“Astronomers have completed a comprehensive examination of large asteroids that zoom near our planet, determining that Earth probably won't be struck by such an object for at least 1,000 years.“ — Only valid for large asteroids of 1 kilometer width and above, so impact of a small space mountain on the CCC headquarters is still very possible. Let a man dream.
I can tell you a batshit story about some dogshit bullshitty horseshit, it's absolutely ratshit and apeshit at the same time, and you never would guess who's the chickenshit dipshits who started this shit. And this is no shit, actually.
GOOD INTERNET is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.