Deviant Slop: Power and Control in AI | Jaya Klara Brekke, Alessandro Longo, Alistair Alexander

Yeah. So, my name is Jada. Uh I have a variety of different hats um on a kind of practical level and part of um the NIM team. uh that's uh uh through the build privacy software and more specifically we've built that kind of overlay network that adds noise to traffic as it kind of goes across the the internet um and uh routes traffic through independently operated um relays which uh prevents patterns of communication from nothing being visible. Uh we've built a kind of like VPN uh interface on top of that.

Um if anyone wants to try that you cann do come up and you know talk to me afterwards. Um we've been working very closely with brave browser resetting the envelope at home connect apper this sort of thing to it. Um so that's kind of like uh let's say the practice side of my work. Um that's you know it's very very important to me um on the kind of war theory side I suppose that's kind of where I know Sana absolutely brilliant fantastic um yeah I've been kind of working on this topic of technology and power for um many many years from a variety of different angles. um my interest in in you know power and how that is distributed was also one of the main reasons why I ended up building privacy technology and getting involved in in this space um and uh but I would say like my more kind of uh uh in more specific interest right now um has to do with the intermediation I suppose between um you know ourselves and our formation of identity um via and through surveillance based technologies um and also in relation to others and so this kind of like production of an algor algorithmic society through this sort of unwitting participation that happens um by engaging with digital technologies that at the same time are um extracting and and expropriating our kind of uh human uh stuff let's say.

So that's the the quick summary I suppose where I'm at both on a on a practical level and and theoretical level.

Alexander.

Hi. Yeah, I'm Alistister. I'm I guess at the moment I'm researching and writing about the social and ecological impact of technology and I guess that's kind of this kind of information environment of deviant slop is creating. So I I guess that brings to the point I would say is that an awful lot of what we're seeing with AI is deviating in a very fundamental way from some of the basics of capitalism which were there in the first place. um I mean in in the sense of it's going so far away from any kind of rational reality and rational connection to what's actually happening in the material environment um that it it has no meaning anymore except to create more and more capital for fewer and fewer people.

So um that's a very roundabout way of getting round to what what what I'm thinking about it and how to approach it. And my my point would be that it's that material reality of our technology that we have to to reconnect with and re understand and and and and learn about and understand that kind of environmental impact. And for my from my work, what I'm trying to do is think about how we can have a radically different relationship with our technology, how we can rebalance our own lives so that we're thinking more about our connection uh to each other uh irrespective of technology. So things like uh intentional analog practice for example um thinking about how we can actually resist those kind of infrastructures. There are huge data centers popping up all over Brandenburg for example right now.

So this is a huge issue for us and we really have to just think about the ways we use these technologies to to find different ways of connecting and different ways of using that technology to to actually bring us back into actually our own physical embodied beings. I think and that's the most important thing which would just kind of be the ultimate antidote to actual AI slop I would say

as a as a yoga teacher. So in one of my many lives I'm also a yoga teacher and that sense of kind of yeah moving between like the body and the breath and then like the daily life is so apparent when you go into a yoga studio and everyone's like oh yeah I'm in my body and then they're back into the into the world of everything all the time. Um, it's interesting that that when you start talking about deviance or when one starts talking about deviance, you automatically start talking about the human in the system or kind of the material reality of the body and and I think that's something we'll come back to in a minute.

Um, but Aleandro would you share kind of your thoughts on the on the panel proposition? Yeah, I mean about the deviant I think I agree with Ja in the sense that I don't think there is any outside from the the main tools through which we distribute cultures today that are algorithmic tools that we don't have any ownership on it but in at the same time I think that in this relationship there's still a space for agency and I think that maybe something that uh the more like privacy maxing the narrative gets kind of wrong is that people are still feeling that they are doing something in these systems and this not only 1984 top down control like that's not the perception of users like people online you know cuz like and this there is this book that I actually haven't read but algorithm of resistance by it bonin and someone else that talks about labor and people like working for delivery services and so on and how they develop a set of practices that like always you know contracting with algorithms so also this kind of paranoid vision of like uh I don't know recommendation algorithm of any kind always being applied uh with like 100% of success like I don't think that this is actually what what is happening right at the same time you have to deal with the fact that every and I think this every like subcultural seed or like scene niche thing like orange wine or sardines or whatever you know it's being like regurgitated by this system you know and I think especially in my work recently I trying to also mark the shift between the the recommendation algorithm of like Facebook and Instagram and then Tik Tok and then Tik Tok and the for you page because the for you page doesn't work like a recommend it's not based on the social graph but it's based on the content recommendation and clustering which is a bit it gives you the perception that you can access to different things without them they don't need to be close to you now I don't want to be it's not a technical explanation clearly but I think we see that in the proliferation of like aesthetics and the course and this like little units of pseudo subcultural expression that is never uh however like really rooted in a in a community or in in in a physical place that is like also how historically subcultures have been developing. uh so I think it's it's not easy to find uh a real deviance and probably the devian always being reabsorbed but also I I would say that that I mean also to connect with what Alistair was saying like on the other end you have this uh there's also a power in like admitting that uh we are all part of the infrastructure and there is no outside but that still people can find protocols of agency I think to move throughout this infrastructure and to still have a certain uh degree of expressivity and and possibility and I think artists should steer this movement you know and should also as I was saying I mean at least that's what I'm interested in trying to see like in Keros Kuro you know in between what what like even expressions that are never purely purely like expression for autonomy or like they're not real autonomy because they're happening in this infrastructure they can contain the seeds for something that could move beyond you know so I think there is always interesting work to do that and maybe that's also the the goal of like I don't know art or theory

I think picking up on that sense of human expression then of maybe like the desire to participate to express and you know it is we're in a visual medium I know that here we're also looking behind the curtain of that visual medium but really the the sharing of memes the sharing of images the collaboration with AI on slop imagery you know like Um, how for you is is then that sense of the visual or aesthetic culture and the traveling of aesthetic culture kind of landing and where are maybe the pockets of kind of I guess like aesthetic resistance.

I can throw that back. Yeah. Anyone pick up now you

uh uh I mean about the lop I think the lop actually was over discussed. I don't think it's so relevant. Uh honestly like the image production like who cares it's just images right and I mean some are just memes then you can say that maybe we are like way further in the process of like loss of meaning because you have a zang tunum and ballerina capuchin and they is not literally means anything but I don't think that AI slop as a visual phenomenon is so interesting visually in terms of power you see the Gaza tower and all that no horrible content And that's a way of like uh I think prefiguring violence or or simply like propaganda in a very similar way how like historical fashions use cinema. I think then for more optful content I like the cyber deck trend. Uh so it's just like this people building uh it's an online trend indeed but it's like people building these mini computers in very personalized shells and so on.

And I think it's interesting to see uh yeah I don't know just people swinging uh and putting together Raspberry APIs and keyboards you know doing this very aer activities uh from a different perspective is usually very female coded like way more than the typical you know aer or yeah custom computing space. So I think that's like kind of calling for another you know a more like brick relationship with technology like more in bolage rather than in engineering. So I I like that. I think that's some aesthetic expression that's interesting.

I I think maybe you know with you also thinking about the sense of the analog. You know we heard earlier about vinyl people coming back to vinyl. So this kind of like sense that people will always go back to um an experience that is kind of deeper or you know with vinyl obviously the sound quality is so much you can hear it your with your ears. There's just such a difference between listening to compressed sound and final sound that you can touch and move. So I think there's a sense of yeah analog creativity being something people still crave because of sensory connection.

And I wonder in your work, you know, when you talk about coming back into into the body in that sense, how how then you're watching these ecosystems evolve of culture. Yeah, I think um I think I think the the the issue is that that it we creates an environment online where it's very hard to distinguish anything which isn't the slob and is from a different position anyway. So I think I think increasingly people are trying to find sort of ways of creating boundaries around particular spaces so they can actually u develop things which are more I think authentically human in different ways. I guess one one aspect of that which I think is really interesting uh is there's a kind of kind of sort of growing sort of and community of people um practicing permacing and the idea is to sort of bring in the whole idea of perma culture into computing which is kind of two quite conflicting kind of ideas but um the idea is really to understand the the materiality of our hardware and understand there's a core sort of belief in care for life care for the chips and really think about the fact that the fundamental microchips on which all of our digital technologies are made must be uh considered as things that need to be preserved and used as much as possible. And furthermore, that older technologies can be used and can be just as good and better in fact at doing things uh as newer technologies.

So like going back and finding older sort of uh games consoles, older computers, older um different types of devices and trying to make them work and in different and simpler ways which are far more aligned with our our own human connection but also also uh our environmental impacts. Um and and that has created a whole different kind of aesthetic and sort of you know much more minimalized forms of digital communication much uh sort of slower forms of digital communication as well and um I think a lot of people find I mean it's quite a niche area I guess but it's uh something that people find uh definitely has a a kind of a more I don't know a better way of of interacting and connecting with other people and with techn technologies and and and and rediscovering some of the sort of initial kind of interest and um I don't know uh inspiration people found in technologies you know 20 30 years ago when the uh when it all seemed sort of relatively exciting uh and and and benign and and it didn't turn into this horrific monster that we got all around us right now. So, I I think there's a whole bunch of different things there and that I think really kind of links in with analog as well to find uh different ways of doing that. And I I just think this kind of idea of creating these boundaries around things and finding different ways of of of actually sort of trying to block out spaces where people sort of there's there's another kind of online space called the the was it is it the the forest OS dark forest OS which is kind of a nice little space where people have created their own social media and is kind of recognizably different and I think these kinds of spaces are becoming more and more interesting to people and friction maxing and this idea of creating more and more ways in which we don't sort of get sucked into this frictionless uh UXoriented environment. I think we need all of those things.

Yeah.

Uh I have lots to say on so many things that have been said just now. I'm trying to think of kind of the right way to tackle it. But um I suppose I mean first of all slop like I think actually um slop played a hu hugely important role in um let's say mainstreaming uh like a certain kind of set of understandings around the image economy and how like AI actually works um for people. So, it kind of I think it helped people understand that AI does hallucinate and that you can't always trust everything that you see and there's like there's been a lot of like value to that in an interesting way and it's opened up a lot of conversations around that. Um, on the flip side, it's also caused a lot of problems because then people don't trust what they see and in certain sit situations they do need to trust what they need to that what they're seeing.

So, you know, in terms of evidence and so on. Um and so there's kind of you know a lot of work to be done and a lot of interesting stuff happening in kind of you know verifiable um uh like architectures of verifiability and and kind of you know obviously cryptography in various ways plays a hugely important technical role there. So you know I think slop has been important in many ways um to kind of popularize like um a lot of cultural theory across the mainstream in a very very quick way um and immediate way. Um, and of course we see wonderful examples of slop from my perspective. There was that uh I don't know if people saw the Lego uh videos that were being made kind of trolling Trump and the US when they were kind of starting the air strikes against Iran.

You know, there's there's interesting fun stuff happening. There's clever ways of engaging with that whole kind of world and that that image economy. Um, but yeah, like I said, kind of slop also produces like noise, right, and informationational noise. And I suppose that's the kind of big problem right now is like how do you actually make sure that the stuff that really matters can get surfaced um in the type of you know infrastructures that we're that we're inhabiting today. Um and that's a far more kind of like complex and important um set of questions.

um when and and I I agree actually with Alistair there that there is this kind of like effort right now where people are trying to build you know some sort of like spaces of trust and verifiability in like their communities in various ways and I think um the kind of technical experimentation there is super interesting. The permacing people um the kind of like rise in popularity of group chats and um spaces that have more of that kind of uh building on the kind of like human trust um element. Um, so, uh, but yeah, I mean, when when when this kind of title of the panel first came up, um, deviant slop and I was trying to think of how my work kind of intersects with it, it kind of does loosely in many ways, but I suppose, you know, playing around with the term a bit creatively. I think it's maybe also what you know on a more symbolic level what we try and do with the kind of nim network. It's like, okay, so the slop can be also understood as noise.

how can noise become a kind of cover um to protect each other and so you know we're adding noise to the traffic you know and that's also working as a kind of anti- AI kind of um infrastructure in a sense you know so yeah there's there's lots of angles there but um yeah

yeah and I think when I you know when I first kind of encountered the term or kind of used the term loosely a while ago I was like oh does it work but I think really thinking in in terms of what you were saying there about truth you you know, the truth of the image, the noise of the image economy, like what is and isn't real. Obviously, we then think of that in the current context of a kind of rise in fascistic thought and action. Um, and I think a lot about Hannah Arent in that context and her warning that, you know, the loss of truth is the point at which authoritarianism and totalitarianism can take control. And so I think there is a much more political angle to this sense of the noise and the slop which I'm pleased that you brought up there.

Yeah, absolutely. I mean Arent is so important to read reread right now.

Yeah,

there's another um uh actually Artent came up um recently in a a wonderful um talk that I saw by Mercedes Boon um that I think might also be relevant for this conversation. But um Arent was talking about technology as being kind of the the you know harnessing of um you know nature in order to kind of channel the power of nature into the human artifice. But what's actually happening with AI is um it's the it's the kind of inverse. It's the channeling of the kind of human into the artifice of the techn of the technological in order for kind of um new forms of exploitation and financialization to take place. And that's something that I that I personally find very interesting because I'm very focused on this kind of the the formation of like identity and by nature of the formation of identity, the formation of relationships between people and the kind of production of social relationships and society um so to speak.

And so I think there's something very very deep that's happening there um when it comes to um that form of the exploitation of of the self and the humanness um and the intermediation in the formation of our own identities um and the formation of our own kind of collectives that is pretty profound very very novel um and really needs like attention and is also where kind of privacy tech can kind of come in and play a bit of a a you dampening role, a bit of protection. Um, but yeah, that's kind of personally where my my interest is right now. Yeah,

thank you so much. We're about to run out of time. Um, but does anyone have any closing reflections just on that one or anything else that we have discussed? Yeah, I just had one thing I was thinking about with this and and you know so to my mind you know I used to work a lot on disinformation actually and a lot of previous projects I did a few years ago with actually a lot of people here with Merrick I did a disinformation project also with Tatiana and Yonas and and that was a really important uh thing to be looking at a few years ago but then increasingly it became clear and I think with with AI and and with this kind of slop then I've moved away from disinformation as an issue. And I I wondered I was reflecting on that recently and I realized that the the problem is that you know when the people who own the platforms are actually pushing the disinformation whether it's SLO or whatever's going on in X.

com now it's like the disinformation is not the problem anymore. You know you can't continue thinking about that as the thing you're focusing on because the kind of people who are owning it are actually pushing it. So the problem is a higher level. So I think when we're looking at these things, we really need to look at the ownership structures of the actual um infrastructure that is actually pushing the stuff out in the first place. That's a really critical element of this.

Yeah. Uh maybe just that exactly. I also think it's more of a power dynamic thing. So it's also important to build counter counter infrastructures and counter power and I think also bringing people together and and realizing how much these technologies are shaping our identities and large language model very rapidly. So the way we work and so on.

So I really think is is important to also develop what I call like tiny protocols but like ways of engaging that are like shared and they can somehow yeah preserve the friction and the the human side of things. So yeah

I think I think about that as well and I'm a writer and sometimes I now think in like you know meme text or like Instagram caption and I'm like oh god that's horrible you know and it's it's in there. Um thank you so much for a fascinating panel. Thank you.