- They're already well able to surface the most addictive short video for a specific user out of millions of real videos.
- But these millions of real videos are just darts thrown into the space of "videos that could hook the user", in the end even the best-selected of them is not perfect.
- Now, behold! AI allows to generate the perfect video to surgically hit all the switches in the viewer's brain and turn it into a zombie hooked for days on end.
Let's hope our regulations hit these "social networks" hard enough so that never dare deploy this kind of technology.
We have not yet seen* the kind of large-scale, individually targeted psychological manipulation that cloud AI products can deliver. And I have no doubt the likes of Dario Amodei and Sam Altman will show us, if we give them enough time.
* I suppose the GPT-4o sycophancy/AI psychosis crisis was a preview, but that was just blunt “engagement” tuning.
This is a tool to help researchers in figuring out what different parts of the brain are actually for with less experimenter bias contamination of “well we think maybe it’s about this so let’s show it video of x to see”.
The essence runs on having someone sit in a scanner for a couple hours watching all sorts of things, and then feeding that to a model that will then build its own representation of said data and try different things on it until it’s found what makes a certain part sing in the model.
The purpose is a generalized understanding of brain function, more or less the same way we’ve been doing it all these years. Expose brain to something, record it somehow, see if brains reaction in the recording helps you understand more about who we are and what cognition is.
(STS seems to be https://en.wikipedia.org/wiki/Superior_temporal_sulcus in the temporal lobe, so I guess its the "I sense a presence" region.)
Explain the potential to exploit strong stimulation of specific visual regions for evil. "Oh, I very much detect a face/place/body/motion/pattern/human", says the subject. What are you going to do with that, startle them?
It also helps companies like Moonbug Entertainment (Candle Media) understand how to build better Distractatrons.
It’s a small TV screen, placed a few feet from the larger one, that plays a continuous loop of banal, real-world scenes — a guy pouring a cup of coffee, someone getting a haircut — each lasting about 20 seconds. Whenever a youngster looks away from the Moonbug show to glimpse the Distractatron, a note is jotted down.
“It’s not mega-interesting, what’s on the Distractatron,” said Maurice Wheeler, who runs the research group. “But if they aren’t fully focused, they might go, ‘Oh, what’s that?’ and kind of drift over. We can see what they’re looking at and the exact moment when they got distracted.”
https://www.newyorker.com/magazine/2024/06/17/cocomelon-chil...
What a world.
I can see it now.. The Distractatrons: a new chapter of protagonists in Transformers! The modern equivalent of evil in this day and age of ADHD and low attention span!
https://en.wikipedia.org/wiki/Looker
the same year as Baudrillard's Simulacra and Simulation, featuring computer generated characters, brainwashing television commercials, a light pulse gun that causes absence seizures, gun battles inside a plastic surgery clinic and an AR simulation environment, a sadistic computer, a physician who doubles as an action hero, and James Coburn giving a lecture explaining the enemy's evil plans in the opposite role that he played in the (excellent) The President's Analyst.
This time I was not so dazzled and saw it for as atrocious everyone else things it is. The minions of "Digital Matrix, Inc." manage several assassinations with the light-pulse L.O.O.K.E.R. gun but when they use real firearms they outdo Vader's stormtroopers by shooting each other. (Want to see the scene where somebody from E.Y. tells them to stick to the L.O.O.K.E.R. gun) The bad guys explain the penultimate secret to the protagonist early on but the ultimate secret is revealed in the L.O.O.K.E.R. lab which doesn't feel like a lab at all but rather a rather good room in a theme park experience where you're supposed to uncover the secret. (Contrast that to the lab Doug Trumbull outfitted in the Brainstorms movie a few years later which is packed with real surplus equipment... I've been to that lab!)
https://en.wikipedia.org/wiki/Supernormal_stimulus
https://en.wikipedia.org/wiki/BLIT_(short_story)
Also: one of the V3A animations reminds me loosely of things I saw when I was a kid, at night, shortly before I slept (though my experience then was more circular).
Also relevant: <https://www.quantamagazine.org/a-math-theory-for-why-people-...>
My understanding is that those who work with the mentally handicapped use bright lights and other stimuli to soothe and control them. It is also my understanding that the autistic are stimulated by vibrant colors (coughcoughMy Little Ponycoughcough).
Who is to say that the rest of us are not also vulnerable to such controlling stimuli?
Or how a very useful tool can become a public health catastrophe.
If future generations of researchers will wonder why IRB reviews became mandatory for computer science, studies like this will be the answer.
Seriously, some people don't seem to realize the point at which they are becoming Fritz Haber.
And I have yet to see a single paper like this where a researcher bails out and publicly says they refuse to work on such projects. Not one.
The most benign interpretation of this observation is that science is filled with spineless opportunists who don’t care who they hurt with what they create. A slightly less benign interpretation might be that many of these people are doing this deliberately, and getting off on the sense of power it gives them.
When it is pushed from the top it is hard to stop at ground level.
Realistically, probably ads, but maybe not only that?
(AI start-up idea: one of our ads a day keeps dementia away! /s)
Without ads and exploitation of the masses, none of these would not be possible.
>Those ancient wonders were enabled by slave labor
That is exactly the point.
What could possibly go wrong?
But for the paper itself, it seems they're using genetic optimization over predefined keywords. Wonder what would happen if they did gradient descent on the latent space directly. Is brain stimulation just not a good domain for GD?
The largest LLMs right now are at best 1% the number of parameters of a human brain.
"At best" if synapses are one parameter each, real ones are probably more than that, but nobody's entirely sure yet.
reason I am asking it could be some relief to our brains after tedious working day, especially after heavy AI usage
if it is targetting visual regions of brain and I have aphantasia (I cannot visualize anything in my mind) is that connected?
It's so cool that you have an MRI handy to check such things.
I don't think it's a matter of if but when. Grim.
But here we can start also the usual discussion about technology research for the sake of it vs calibration of possible side effects of new research
Personally i think we haven’t solve this problem and thus it’s just a matter of time until we’ll get in a non-going-back point
I wish we had the Hippocratic oath for STEM, or at least that they would take ethics seriously rather than an afterthought against the god of Progress at all costs.
Will be interesting to see how strong the controlling forces can be - enough to make you miss things in direct perception like in the book, or only softer effects further up the cognition layer stack
Nothing special compared to purpose made screen savers.
Half joking, half paranoid.
Is my brain different or am I just a grumpy millenial hipster?
All of this to say, if you subjected yourself to just enough TikTok scrolling on just the right topic, you might find yourself using it occasionally after that initial hump, then slightly more frequently, then daily.
You might still not "like" it, but the habit is what matters.
Prior exposure to worse feeds gives like an analytical look on the vids rather than emotional. I am fast scanning for the joke. Or something.
I do enjoy watching YouTube videos at home, on the living-room flatscreen, on a variety of topics, but I select them manually, one at a time, from the vast selection The Algorithm(TM) offers me, plus my own searches.
The fact it’s bucketing by making images of lighting and facial expressions, the fact it doesn’t natively do the video it does an image then video generates from it.
The results look really bad and samey. Doubt this would work for the actual thing they’re pitching it for.