Re: Spamming the Data Space – CLIP, GPT and synthetic data

[jan hendrik brueggemeier]

For what it is worth, I would suggest to keep “spamming” in the title of 
your paper, Francis. It it a value judgment but I would assume the whole 
point it to call out BS when you see it - flood like quantities of 
images generated by positive feedback loops fueled by “biased” 
algorithms written for ROI does call for an interrogation, doesn’t it? 
The question of purity of culture is bit of a furphy in my view because 
even the the Bitnik’s Dada piece kicks off their google search image  
with The Poet and, well, you would hope Man Ray got paid for making it. 
So culture/commerce/capitalism then becomes about the initially 
mechanistic and today more processing scale of distribution. I suppose 
the issue is more than of the online dissemination of the images and how 
they, for instance, interface with search engines or other databases etc 
… where do they all end up? I guess this would bring us back to the 
spamming attribute.


To add a blow from the past -- as this thread made me think of Vilém 
Flusser’s techno images, if I am not mistaken, he even spoke of 
dialogical images in the context of digital technology (and networked 
dissemination). Revisiting over the last couple of days his work from 
the 90ies, I find it fascinating to go over the "our live in houses that 
have become ruins" phenomenological speaking as communication tech 
tearing wholes into them and helping us to dialogue with each other (and 
this way we may even communicate our way out of nationalism etc). Well, 
I guess it was the visionary aspect of the 90ies but it also links to 
Brian's point "with statistically generated images you are in a sense 
alone in the room, there is no one to evaluate or answer to" and 
possible Geert's extinction internet question if an alternative (aka 
non-commercial) internet culture should have more on offer than FLOSS 
values but not necessarily no alternative social experience (other than 
platform social media as we came to know them so far). So are we now 
talking more than just corporate capture of platforms but cognitive 
human limits? The former could be considered fixed with a change over to 
mastodon (and I am enjoying my random visits -- and I would like to 
thank nettime for the prompt -- but I also currently in the privileged 
position not necessarily depending on an established personal brand to 
make a living ...) but to cut to chase here would the question remain: 
is this form of social media need other forms of restraints? Is it still 
too 24/7? Still too "individualistic"? Too resource-hungry etc?


Anyway, my new 2 cents & happy new year everyone! ;)

jan

On 31/12/2022 7:10 am, Francis Hunger wrote:


Hi Luke, hi everyone

first of all thanksto everyone for your reactions,
Why is my DALL-E generated image 'derivative' or a 'tainted' image, 
according to the tech commentators I mentioned earlier, and my 
'manual' pixel art not? I honestly don't see what the difference is 
between a MidJourney collage and a Photoshop collage. The same 
question goes for text. Why is Hardy Boys #44 (or whatever formulaic 
fiction you like) a 'real' book and the same story cranked out by 
GPT-3 'contaminated' or 'synthetic'?


Lot's of these differentiations plays between 'human' and 'machine' 
creativity plays out beyond the background of AI phantasms. That is 
the phantasm of full automation and of machine domination. Let me 
self-quote from a text we published earlier this year


"AI contains multiple phantasmatic narratives. First, it can be said 
that it masks human fear of death by imagining a possible continued 
life as a machine (in the transhumanist movement). Second, it 
constructs AI as ‘the other’ of hu- mankind. This phantasm draws on 
people’s longing to be relieved from labour, for example, by digital 
assistants coordinating their schedules or by autonomous cars.
Further, the ‘other’ of humankind is reflected in the fear that humans 
could be overwhelmed by the machine developing a kind of ‘super 
intelligence’.6 It is present, for example, in numerous movies and 
science fiction novels in which AI is depicted as humanoid robots. In 
this phantasm, humans are positioned as ‘the natural’, ‘the 
primordial’, and the machine is ‘the artificial’ to be distrusted.
It is not only fears, but also desire that is linked to the phantasm. 
The digital assistants that free us from labour portray the desire for 
freedom from the yoke of wage labour to which people in capitalist 
societies submit. The AIs, on the other hand, which take over the 
world, as in the films Ex Machina (Garland 2014) or Free Guy (Levy 
2021), visualize the actually inexpressible wish for submission, 
which, similar to a sadomasochistic relationship, also means the 
freedom of the submissive, namely the freedom from responsibility." 
(Arns, Hunger, Lechner 2022:34f, 
https://www.hmkv.de/files/hmkv/ausstellungen/2022/HOMI/Publikation/House%20Of%20Mirrors%20Magazin%20PDF.pdf)


In that sense lot's of the current discussions

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Francis Hunger]

Hi Luke, hi everyone

first of all thanksto everyone for your reactions,
Why is my DALL-E generated image 'derivative' or a 'tainted' image, 
according to the tech commentators I mentioned earlier, and my 
'manual' pixel art not? I honestly don't see what the difference is 
between a MidJourney collage and a Photoshop collage. The same 
question goes for text. Why is Hardy Boys #44 (or whatever formulaic 
fiction you like) a 'real' book and the same story cranked out by 
GPT-3 'contaminated' or 'synthetic'?


Lot's of these differentiations plays between 'human' and 'machine' 
creativity plays out beyond the background of AI phantasms. That is the 
phantasm of full automation and of machine domination. Let me self-quote 
from a text we published earlier this year


"AI contains multiple phantasmatic narratives. First, it can be said 
that it masks human fear of death by imagining a possible continued life 
as a machine (in the transhumanist movement). Second, it constructs AI 
as ‘the other’ of hu- mankind. This phantasm draws on people’s longing 
to be relieved from labour, for example, by digital assistants 
coordinating their schedules or by autonomous cars.
Further, the ‘other’ of humankind is reflected in the fear that humans 
could be overwhelmed by the machine developing a kind of ‘super 
intelligence’.6 It is present, for example, in numerous movies and 
science fiction novels in which AI is depicted as humanoid robots. In 
this phantasm, humans are positioned as ‘the natural’, ‘the primordial’, 
and the machine is ‘the artificial’ to be distrusted.
It is not only fears, but also desire that is linked to the phantasm. 
The digital assistants that free us from labour portray the desire for 
freedom from the yoke of wage labour to which people in capitalist 
societies submit. The AIs, on the other hand, which take over the world, 
as in the films Ex Machina (Garland 2014) or Free Guy (Levy 2021), 
visualize the actually inexpressible wish for submission, which, similar 
to a sadomasochistic relationship, also means the freedom of the 
submissive, namely the freedom from responsibility." (Arns, Hunger, 
Lechner 2022:34f, 
https://www.hmkv.de/files/hmkv/ausstellungen/2022/HOMI/Publikation/House%20Of%20Mirrors%20Magazin%20PDF.pdf)


In that sense lot's of the current discussions play out. And even when 
trying to avoid it, obviously the /Spamming the Data Space/ enables 
phantasmatic readings and also builds on the automation phantasm.


So the point is more about creating data loops (and maybe I should 
change the text's title towards /Data Loops/), as for instance this 
recent paper shows: Too Good to Be True: Bots and Bad Data From 
Mechanical Turk https://journals.sagepub.com/doi/10.1177/17456916221120027


Many of these posts have suggested future autonomous zones where 
'synthetic' culture is banned. What would be the hallmark or signature 
of these spaces? No digital tools or algorithmic media may come to 
mind, but these overlook the most crucial element to 'new' cultural 
production: reading or listening or viewing other people's works.


When I allude back to the no-photo policy in clubs, basically upon 
entering said club you get a friendly reminder to not use your mobile 
phone for taking pictures, or you get a sticker put onto the camera 
lense and it is clear to everyone, that this a a rule to which you 
_want_ to comply because you want to create a common space. So, 
certainly we are talking more about a social approach and less about a 
technical.


best

Francis




- 'Perplexed in Mianjin/Brisbane'




On Fri, Dec 23, 2022 at 8:54 AM Francis Hunger
 wrote:

Dear Luke, dear All

Interesting essay Francis, and always appreciate Brian's
thoughtful comments. I think the historical angle Brian is
pointing towards is important as a way to push against the
claims of AI models as somehow entirely new or revolutionary.

In particular, I want to push back against this idea that
this is the last 'pure' cultural snapshot available to AI
models, that future harvesting will be 'tainted' by automated
content.


At no point did I allude to the 'pureness' of a cultural
snapshot, as you suggest. Why should I? I was discussing this
from a material perspective, where data for training diffusion
models becomes the statistical material to inform these
models. This data has never been 'pure'. I used the
distinction of uncontaminated/contaminated to show the
difference between a training process for machine learning
which builds on an snapshot, that is still uncontaminated by
the outputs of CLIP or GPT and one which includes generated
text and images using this techique on a large scale.

It is obvious, but maybe I should have made it more clear,
that the training data in itself is already far from pure.
Honestly I'm a bit shoc

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Luke Munn]

Yes thanks to Brian and to Prem for their beautiful and thought-provoking
responses to my questions.

To briefly bring it back to the starting point - I think the ambiguity I
was pointing to is that these AI models blur the boundaries of some of
these traditional distinctions. They use very 'rational' forms of
statistical inference for instance, but they do so to make connections
between items in latent space (e.g. 'relationships'), and their outputs can
be very funny, cute, shocking, and playful or otherwise affectively dense.
All of this *feels* quite different from the cold, logical associations to
'intelligence' that computation has traditionally had. I've written a
little about this alternative strain of computation, and recent attempts at
becoming more empathically attuned, in my book 'Logic of Feeling'.

Some of this AI 'progress' is a result of recent architectures (e.g.
Transformer models). But some of it comes from harvesting the social
en-masse in image or text form. And some of it comes from introducing new
layers of feedback and fine-tuning from human workers ('reinforcement
learning through human feedback'). These layers and circuits of sociality
make it very hard to pinpoint where the 'human' ends and the 'machine'
begins. And this slipperiness then pervades the resulting outputs, in
Prem's words, 'to the point where we can no longer differentiate the
human'.

Already we've seen AI outputs win art competitions. And distinguishing a
'real' student essay from the AI generated one is now a lost cause. These
are banal examples and again, this is not to lionize machine learning
models or Big Tech, but simply to point to the ambivalence that these
technologies open up. So while I agree in principle that 'synthetic images
have no punctum', I really couldn't establish some test that distinguished
'synthetic' images from 'authentic' ones.

But these more particular points I think overlook the larger and more
fundamental ideas in Brian and Prem's posts. The atomizing,
hyper-individualizing tendencies in neoliberalism. The narrow, impoverished
understanding of 'rational' intelligence in the West. And really, the
growing sense that these dominant forms of knowing and being are inadequate
for our contemporary conditions and planetary crises. A better attunement
to the environment, solidarity with forms of productive and reproductive
labor, a radical expansion of what counts as 'knowledge'---all of these
will quickly move from 'virtue signalling' to fundamental survival
responses over the next two decades.

ngā mihi / best, L


On Thu, 29 Dec 2022 at 08:38, Brian Holmes 
wrote:

> Prem, your last post is spotlight and lantern all at once: extremely
> precise, diffusely warm and wise. Thank you very much.
>
> Large language models don't understand anything. Synthetic images have no
> punctum, no trace of vulnerability or mortality. I'll take nettime over
> ChatGPT any day!
>
> Brian
>
> On Sun, Dec 25, 2022, 01:43 Prem Chandavarkar  wrote:
>
>> The true question is how we recognise the other, and perhaps the fault
>> lies in our assuming we do it through intelligence. As neuroscientist, Anil
>> Seth, observes, we hear a lot of talk on artificial intelligence but never
>> hear anyone speak of artificial consciousness. And that is because
>> consciousness is tied to being a living, breathing, embodied being, whereas
>> intelligence, because it lends itself to abstraction, does not suffer this
>> constraint to the same degree.
>>
>> In his book “The Master and His Emissary: The Divided Brain and the
>> Making of the Western World,” Ian McGilchrist refers to the popular myth
>> that we have a brain divided into two hemispheres because each hemisphere
>> performs a specialist function, with the left brain tackling the logical
>> subjects of math and language and the right brain tackling creative
>> subjects like art. While neuroscience has rejected this myth decades ago,
>> it still survives in the popular imagination. Both hemispheres tackle the
>> same subjects, but in different ways: the left brain is detail oriented and
>> the right brain is context oriented. McGilchrist argues that this is a
>> primal quality springing from our evolution: when we were hunter-gatherers
>> we needed a detailed attention to the world to capture prey or gather food,
>> while at the same time we needed a contextual awareness of the world to
>> watch out for predators. As per McGilchrist, our notion of modernity has
>> been shaped by Western civilisation, with the Enlightenment privileging
>> instrumental reason as the foundation for democracy, given that everyone,
>> irrespective of their birth, has the capacity to reason. The institutions
>> of modernity operate on protocols predicated on reason and discourse, and
>> therefore we neglect our consciousness that contextualises us within the
>> world.
>>
>> Alison Gopnik, in her book “The Philosophical Baby,” comes up with a nice
>> term for these two types of consciousness: spot

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Brian Holmes]

Prem, your last post is spotlight and lantern all at once: extremely
precise, diffusely warm and wise. Thank you very much.

Large language models don't understand anything. Synthetic images have no
punctum, no trace of vulnerability or mortality. I'll take nettime over
ChatGPT any day!

Brian

On Sun, Dec 25, 2022, 01:43 Prem Chandavarkar  wrote:

> The true question is how we recognise the other, and perhaps the fault
> lies in our assuming we do it through intelligence. As neuroscientist, Anil
> Seth, observes, we hear a lot of talk on artificial intelligence but never
> hear anyone speak of artificial consciousness. And that is because
> consciousness is tied to being a living, breathing, embodied being, whereas
> intelligence, because it lends itself to abstraction, does not suffer this
> constraint to the same degree.
>
> In his book “The Master and His Emissary: The Divided Brain and the Making
> of the Western World,” Ian McGilchrist refers to the popular myth that we
> have a brain divided into two hemispheres because each hemisphere performs
> a specialist function, with the left brain tackling the logical subjects of
> math and language and the right brain tackling creative subjects like art.
> While neuroscience has rejected this myth decades ago, it still survives in
> the popular imagination. Both hemispheres tackle the same subjects, but in
> different ways: the left brain is detail oriented and the right brain is
> context oriented. McGilchrist argues that this is a primal quality
> springing from our evolution: when we were hunter-gatherers we needed a
> detailed attention to the world to capture prey or gather food, while at
> the same time we needed a contextual awareness of the world to watch out
> for predators. As per McGilchrist, our notion of modernity has been shaped
> by Western civilisation, with the Enlightenment privileging instrumental
> reason as the foundation for democracy, given that everyone, irrespective
> of their birth, has the capacity to reason. The institutions of modernity
> operate on protocols predicated on reason and discourse, and therefore we
> neglect our consciousness that contextualises us within the world.
>
> Alison Gopnik, in her book “The Philosophical Baby,” comes up with a nice
> term for these two types of consciousness: spotlight consciousness (focus
> on details) and lantern consciousness (contextual awareness). Spotlight
> consciousness privileges our own agency, seeking to manipulate the world.
> Whereas lantern consciousness reverses agency, granting it to the world and
> according recognition to its capacity to act on us. Empathy, compassion,
> care, wonder, and so many qualities that make life worthwhile, are
> primarily handled and shaped by lantern consciousness. Spotlight
> consciousness pushes us to detach from the world, lantern consciousness
> pushes us toward immersion in it.
>
> More significantly, the two modes of consciousness place different
> emphasis in the methodologies for developing and refining them. Spotlight
> consciousness emphasises abstraction, intelligence and reason, whereas
> lantern consciousness depends more on embodied and experiential practice.
> Perhaps Seth’s discussion on the limits to artificial consciousness apply
> more to lantern consciousness.
>
> Modern education schools us in spotlight consciousness, but in everyday
> life we intuitively rely so much on lantern consciousness. Take the example
> of friendship. If we sought to find friends through a philosophy or
> rationalisation of friendship, we would have few or no friends. We find
> friends through shared embodied experiences investing in time that opens up
> our lantern consciousness to them, acknowledging their agency, revelling in
> the mutual resonances we discover through serendipity; and soon friendship,
> that was absent in our first meeting, emerges to form the fundamental core
> of our shared experience. Lantern consciousness privileges harmony to
> embrace serendipity, complexity, and emergence. Spotlight consciousness
> privileges understanding to enforce simplicity on a complex world,
> consequently tending toward violence.
>
> Lantern consciousness also grants recognition and agency to nature and
> things, not just to people. Jane Bennett, in “Vibrant Matter: A Political
> Ecology of Things,” starts with Bruno Latour’s critique that the
> fundamental error of modernity, as defined in the Enlightenment, lies in
> assuming that we are the only sentient beings in a largely insentient
> world, and argues that the sentience of nature and things is revealed in a
> recalcitrance that becomes evident on considering longer time scales. We
> need to pivot away from the Enlightenment model and recast our politics
> accordingly.
>
> The limits to AI can be recognised only by acknowledging the limits to
> intelligence itself. We must incorporate in our practices what
> consciousness, especially lantern consciousness, has to offer us. Without
> this che

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Prem Chandavarkar]

The true question is how we recognise the other, and perhaps the fault lies in 
our assuming we do it through intelligence. As neuroscientist, Anil Seth, 
observes, we hear a lot of talk on artificial intelligence but never hear 
anyone speak of artificial consciousness. And that is because consciousness is 
tied to being a living, breathing, embodied being, whereas intelligence, 
because it lends itself to abstraction, does not suffer this constraint to the 
same degree.

In his book “The Master and His Emissary: The Divided Brain and the Making of 
the Western World,” Ian McGilchrist refers to the popular myth that we have a 
brain divided into two hemispheres because each hemisphere performs a 
specialist function, with the left brain tackling the logical subjects of math 
and language and the right brain tackling creative subjects like art. While 
neuroscience has rejected this myth decades ago, it still survives in the 
popular imagination. Both hemispheres tackle the same subjects, but in 
different ways: the left brain is detail oriented and the right brain is 
context oriented. McGilchrist argues that this is a primal quality springing 
from our evolution: when we were hunter-gatherers we needed a detailed 
attention to the world to capture prey or gather food, while at the same time 
we needed a contextual awareness of the world to watch out for predators. As 
per McGilchrist, our notion of modernity has been shaped by Western 
civilisation, with the Enlightenment privileging instrumental reason as the 
foundation for democracy, given that everyone, irrespective of their birth, has 
the capacity to reason. The institutions of modernity operate on protocols 
predicated on reason and discourse, and therefore we neglect our consciousness 
that contextualises us within the world.

Alison Gopnik, in her book “The Philosophical Baby,” comes up with a nice term 
for these two types of consciousness: spotlight consciousness (focus on 
details) and lantern consciousness (contextual awareness). Spotlight 
consciousness privileges our own agency, seeking to manipulate the world. 
Whereas lantern consciousness reverses agency, granting it to the world and 
according recognition to its capacity to act on us. Empathy, compassion, care, 
wonder, and so many qualities that make life worthwhile, are primarily handled 
and shaped by lantern consciousness. Spotlight consciousness pushes us to 
detach from the world, lantern consciousness pushes us toward immersion in it.

More significantly, the two modes of consciousness place different emphasis in 
the methodologies for developing and refining them. Spotlight consciousness 
emphasises abstraction, intelligence and reason, whereas lantern consciousness 
depends more on embodied and experiential practice. Perhaps Seth’s discussion 
on the limits to artificial consciousness apply more to lantern consciousness.

Modern education schools us in spotlight consciousness, but in everyday life we 
intuitively rely so much on lantern consciousness. Take the example of 
friendship. If we sought to find friends through a philosophy or 
rationalisation of friendship, we would have few or no friends. We find friends 
through shared embodied experiences investing in time that opens up our lantern 
consciousness to them, acknowledging their agency, revelling in the mutual 
resonances we discover through serendipity; and soon friendship, that was 
absent in our first meeting, emerges to form the fundamental core of our shared 
experience. Lantern consciousness privileges harmony to embrace serendipity, 
complexity, and emergence. Spotlight consciousness privileges understanding to 
enforce simplicity on a complex world, consequently tending toward violence.

Lantern consciousness also grants recognition and agency to nature and things, 
not just to people. Jane Bennett, in “Vibrant Matter: A Political Ecology of 
Things,” starts with Bruno Latour’s critique that the fundamental error of 
modernity, as defined in the Enlightenment, lies in assuming that we are the 
only sentient beings in a largely insentient world, and argues that the 
sentience of nature and things is revealed in a recalcitrance that becomes 
evident on considering longer time scales. We need to pivot away from the 
Enlightenment model and recast our politics accordingly.

The limits to AI can be recognised only by acknowledging the limits to 
intelligence itself. We must incorporate in our practices what consciousness, 
especially lantern consciousness, has to offer us. Without this check, 
intelligence can, and has, exponentially spin off into territories of violent 
distortion, even more so once the data space becomes contaminated with the 
products of AI to the degree where we can no longer differentiate the human.

Lantern consciousness resists intelligence’s obsession with rationalisation and 
definition. Its reliance on embodied practice recognises that there is no 
stepping away from our primordial roots i

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Brian Holmes]

On Fri, Dec 23, 2022, Luke Munn wrote:

> At the core of all this, I think, is the instinct that there's something
> unique about 'human' cultural production. [snip...] Terms like 'meaning',
> or 'intention', or 'autonomy' gesture to this desire, this hunch that
> something will be lost, that some ground will be ceded with the move to AI
> image models, large language models, and so on.
>

These are old (maybe antiquated?) problems that were central to Continental
philosophy from Heiddeger to Gadamer, Levinas, Baudrillard and many others.
Basically the questions are, Who am I and how do I guide my action amid a
flood of normalizing or coercive cultural contents? How do I know and
recognize the Other in his/her/their full otherness?

As time goes by I have got more interested in Gadamer's focus on
interpretation as the process whereby an individual or community sets their
ethical/political course with respect to the expressions and actions of
others. That will always be necessary in any society - exactly because
there is no reliable benchmark, no fully original expression, no pre-given
authentic self - so the process of interpretation becomes a creative and
always provisional act. However, with statistically generated images you
are in a sense alone in the room, there is no one to evaluate or answer to.
Baudrillard has a great quote on this, which I used in my work on
Guattari's Schizoanalytic Cartographies:

"This is our destiny, subjected to opinion polls, information, publicity,
statistics: constantly confronted with the anticipated statistical
verification of our behavior, absorbed by this permanent refraction of our
least movements, we are no longer confronted with our own will. We are no
longer even alienated, because for that it is necessary for the subject to
be divided in itself, confronted with the other, contradictory. Now, where
there is no other, the scene of the other, like that of politics and of
society, has disappeared. Each individual is forced despite himself into
the undivided coherency of statistics. There is in this a positive
absorption into the transparency of computers, which is something worse
than alienation."

Now, AI brings a new twist to all this: computers are no longer
transparent, we don't exactly know how neural networks function. Like Harun
Farocki in his explorations of machine vision, some people are now
interpreting the expressions of the inscrutable AIs. There's a chance that
humans will learn something fundamental about the potentials of their own
intelligence through this process. However, it is equally or far more
likely that entire populations will be massively confronted with
statistical transforms of previous generations of statistically generated
images, in the scenario that Francis outlines. What's more, it's
exceedingly likely that the whole process of statistical image production
will be carried on coercively by states and corporations, whose intentions
will be masked by the statistical operations. The Baudrillardean worst-case
is getting a lot closer to fulfillment.

I would be glad to learn different perspectives on all this. It's why I
joined this thread.

All the best, Brian





>>
>> On Fri, Dec 23, 2022 at 8:54 AM Francis Hunger <
>> francis.hun...@irmielin.org> wrote:
>>
>>> Dear Luke, dear All
>>>
>>> Interesting essay Francis, and always appreciate Brian's thoughtful
>>> comments. I think the historical angle Brian is pointing towards is
>>> important as a way to push against the claims of AI models as somehow
>>> entirely new or revolutionary.
>>>
>>> In particular, I want to push back against this idea that this is the
>>> last 'pure' cultural snapshot available to AI models, that future
>>> harvesting will be 'tainted' by automated content.
>>>
>>> At no point did I allude to the 'pureness' of a cultural snapshot, as
>>> you suggest. Why should I? I was discussing this from a material
>>> perspective, where data for training diffusion models becomes the
>>> statistical material to inform these models. This data has never been
>>> 'pure'. I used the distinction of uncontaminated/contaminated to show the
>>> difference between a training process for machine learning which builds on
>>> an snapshot, that is still uncontaminated by the outputs of CLIP or GPT and
>>> one which includes generated text and images using this techique on a large
>>> scale.
>>>
>>> It is obvious, but maybe I should have made it more clear, that the
>>> training data in itself is already far from pure. Honestly I'm a bit
>>> shocked, you would suggest I'd come up with a nostalgic argument about
>>> purity.
>>>
>>> Francis' examples of hip hop and dnb culture, with sampling at their
>>> heart, already starts to point to the problems with this statement. Culture
>>> has always been a project of cutting and splicing, appropriating,
>>> transforming, and remaking existing material. It's funny that AI
>>> commentators like Gary Marcus talk about GPT-3 as the 'king of pastiche'.

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Luke Munn]

> I question whether the notion of pastiche makes any sense at all without
> interpretation (preumably Luke has something to say about that). What's
> certain is that the autonomy question is becoming urgent.
>
>
If an AI model produces a tasteless, derivative image in a forest and
there's no human there, is it pastiche? ;-)

Sure, you're right, pastiche is everywhere, and is very much in the eye of
the beholder. Even without direct AI production of images, a lot of the
shows on Netflix, for example, feel strangely familiar, precisely because
they cut-and-paste elements of 'cult' or 'successful' shows together.
'Stranger Things' was essentially engineered from viewer data, and Netflix
knew it was going to be successful even before it launched.

And this kinda brings me back to some of the genuine questions I have
around AI models and cultural critique - the lines here seem really
arbitrary. Why is my DALL-E generated image 'derivative' or a 'tainted'
image, according to the tech commentators I mentioned earlier, and my
'manual' pixel art not? I honestly don't see what the difference is between
a MidJourney collage and a Photoshop collage. The same question goes for
text. Why is Hardy Boys #44 (or whatever formulaic fiction you like) a
'real' book and the same story cranked out by GPT-3 'contaminated' or
'synthetic'?

Molly, in her excellent post, raises the issue of 'false production'. But I
genuinely don't know what that would look like. Is it false because it's
based on existing images (but collage artists already do this), or because
there's not enough 'intentionality' (prompting feels too easy?), or because
it's generated too quickly? In some ways, the exact same critique of AI
production could be levelled at meme creation - choose your base image, add
the classic meme typeface, and pump out a million permutations - but these
images are somehow considered an 'organic' part of the internet, while what
comes next is going to be artificial, mass-produced, spam, synthetic, and
so on.

At the core of all this, I think, is the instinct that there's something
unique about 'human' cultural production. (Even though AI models are
absolutely based on human labor, designs, illustrations, books, etc - their
packaging and obfuscation of this data makes them 'machinic' in the popular
consciousness.) Terms like 'meaning', or 'intention', or 'autonomy' gesture
to this desire, this hunch that something will be lost, that some ground
will be ceded with the move to AI image models, large language models, and
so on. I'm sympathetic to this - and don't want to come across as an
apologist for Big Tech, Open AI, etc. But I guess my struggle is to put my
finger on what that 'special something' is. Many of these posts have
suggested future autonomous zones where 'synthetic' culture is banned. What
would be the hallmark or signature of these spaces? No digital tools or
algorithmic media may come to mind, but these overlook the most crucial
element to 'new' cultural production: reading or listening or viewing other
people's works.

- 'Perplexed in Mianjin/Brisbane'



>
>
>
> On Fri, Dec 23, 2022 at 8:54 AM Francis Hunger <
> francis.hun...@irmielin.org> wrote:
>
>> Dear Luke, dear All
>>
>> Interesting essay Francis, and always appreciate Brian's thoughtful
>> comments. I think the historical angle Brian is pointing towards is
>> important as a way to push against the claims of AI models as somehow
>> entirely new or revolutionary.
>>
>> In particular, I want to push back against this idea that this is the
>> last 'pure' cultural snapshot available to AI models, that future
>> harvesting will be 'tainted' by automated content.
>>
>> At no point did I allude to the 'pureness' of a cultural snapshot, as you
>> suggest. Why should I? I was discussing this from a material perspective,
>> where data for training diffusion models becomes the statistical material
>> to inform these models. This data has never been 'pure'. I used the
>> distinction of uncontaminated/contaminated to show the difference between a
>> training process for machine learning which builds on an snapshot, that is
>> still uncontaminated by the outputs of CLIP or GPT and one which includes
>> generated text and images using this techique on a large scale.
>>
>> It is obvious, but maybe I should have made it more clear, that the
>> training data in itself is already far from pure. Honestly I'm a bit
>> shocked, you would suggest I'd come up with a nostalgic argument about
>> purity.
>>
>> Francis' examples of hip hop and dnb culture, with sampling at their
>> heart, already starts to point to the problems with this statement. Culture
>> has always been a project of cutting and splicing, appropriating,
>> transforming, and remaking existing material. It's funny that AI
>> commentators like Gary Marcus talk about GPT-3 as the 'king of pastiche'.
>> Pastiche is what culture does. Indeed, we have whole genres (the romance
>> novel, the murder mystery, etc) that are ab

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Molly Hankwitz]

Thank you, Francis for this very interesting work, and for the responses from all. Some of you may not know me at all, so by way of introduction, I am an independent researcher/writer/feminist artist working between film, media and cultural studies on history and theory of new media in the context of post-war arts culture.  I have been teaching about AI, AL, and varied artists works in these fields for a few years with great internet as a new medium (AI) comes in. Some salient features of this analysis and the varied responses stand out for further critique. If I am correctly reading your text, Francis, and I thank you Francis for the separation of the two modi, on the one hand there is the potential for new forms of data sets (as assets and commodities) which are “walled” and controlled - as collections or arrays - which will undoubtedly become, indeed already are being made:  Some examples here: Lev Manovich’s PhotoTime projects, Trevor Paglen’s experimental arrays in which he inputs the dataset and then works with it; or experiments with MS-made Genesis Maps software used on/with the METs digital collections. Some of you may have seen or been at this event written about here: https://www.metmuseum.org/perspectives/articles/2019/2/artificial-intelligence-machine-learning-art-authorshipFrancis writes of something similar - controlled data spaces (like online communities maybe, where contributing or donating is controlled by membership?) “We are going to create separate data ecologies, which prohibit spamming the data space. These would be spaces, comparable to the no-photo-policy in clubs like Berghain or IFZ with a no-synthetics policy. While vast areas of the information space may be indeed flooded, these would be valuable zones of cultural exchange.” On the other hand, there is the ability for AIs to remake data based on anything it is allowed to pull in, and to then feed that invention back into data pools and streams. So the question arises about who and how data will be controlled. This question is familiar. Capitalist interests will have their reasons and methods - maybe what Brian alluded to about understanding that we have seen these mechanisms at work before…and there is the decided problem of having capitalist interests further empowered through the accelerated process of reproduction and distribution of it’s wares, especially when this uncontrolled reproduction/replication is prone to prejudices, biases, falsehood, and omissions. Let’s see what we can foresee. Can we foresee cultural progress continuing as a growing alienation among larger sectors of global populations, no longer able to find any useful or rational meaning in a plethora of false production? (- a great resignation from the virtual spectacle and an end to capitalism of this kind - literally it’s collapse from over extension) Will there be regulation of AI in the net, further regulating the circulation of information? Will  those with the least access, those most vulnerable to data mining of their lives, become even more susceptible to exploitation, as they fail to understand how to navigate this new synthetic reality to their own advantage or where self-interest is a manufactured byproduct of the spectacle? Whatever we might foresee, the bandaid of tweaking SEO for the sake of a false surface of BIPOC to add in is lame (Francis’ example about a few more pictures of BIPOC being worked in) Molly On Dec 23, 2022, at 12:48 PM, Luke Munn  wrote:Hey Francis,Thanks for your response. Just briefly I think there was some misframing of my response. It was not meant as a takedown or critique of your text, but rather a more general response to this idea tabled in the opening line, For the last time in human history the cultural-data space has not been contaminated. This is an idea I had seen a few times from others the same day. 'The internet is now forever contaminated with images made by AI,” Mike Cook, an AI researcher at King’s College London' . 'How AI-Generated Text Is Poisoning the Internet' etc. I was kicking around this idea and seeing what it implied, hence delving into this idea of post-truth authenticity etc which your text doesn't talk about. I know you get this stuff, and can see issues with some of these assumptions, which is precisely why I mentioned your dnb example as culture of remixing. Nevertheless I did think your original piece (and most people's work including mine) can be strengthened by thinking longer term / historically, as the presentist framing of tech is often very powerful. That was my only real feedback - but i can see how this more general 'essay as jumping off point' discussion could be misread, so apologies for that. -LOn Sat, 24 Dec 2022, 01:03 Francis Hunger,  wrote:
  

  
  
Dear Luke, dear All

  
  Interesting essay Francis, and always appreciate
Brian's thoughtful comments. I think the historical angle Brian
is pointing towards 

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Luke Munn]

Hey Francis,

Thanks for your response. Just briefly I think there was some misframing of
my response. It was not meant as a takedown or critique of your text, but
rather a more general response to this idea tabled in the opening line,

For the last time in human history the cultural-data space has not been
contaminated.

This is an idea I had seen a few times from others the same day. 'The
internet is now forever contaminated with images made by AI,” Mike Cook, an
AI researcher at King’s College London' . 'How AI-Generated Text Is
Poisoning the Internet' etc.

I was kicking around this idea and seeing what it implied, hence delving
into this idea of post-truth authenticity etc which your text doesn't talk
about.

I know you get this stuff, and can see issues with some of these
assumptions, which is precisely why I mentioned your dnb example as culture
of remixing.

Nevertheless I did think your original piece (and most people's work
including mine) can be strengthened by thinking longer term / historically,
as the presentist framing of tech is often very powerful. That was my only
real feedback - but i can see how this more general 'essay as jumping off
point' discussion could be misread, so apologies for that.

-L

On Sat, 24 Dec 2022, 01:03 Francis Hunger, 
wrote:

> Dear Luke, dear All
>
> Interesting essay Francis, and always appreciate Brian's thoughtful
> comments. I think the historical angle Brian is pointing towards is
> important as a way to push against the claims of AI models as somehow
> entirely new or revolutionary.
>
> In particular, I want to push back against this idea that this is the last
> 'pure' cultural snapshot available to AI models, that future harvesting
> will be 'tainted' by automated content.
>
> At no point did I allude to the 'pureness' of a cultural snapshot, as you
> suggest. Why should I? I was discussing this from a material perspective,
> where data for training diffusion models becomes the statistical material
> to inform these models. This data has never been 'pure'. I used the
> distinction of uncontaminated/contaminated to show the difference between a
> training process for machine learning which builds on an snapshot, that is
> still uncontaminated by the outputs of CLIP or GPT and one which includes
> generated text and images using this techique on a large scale.
>
> It is obvious, but maybe I should have made it more clear, that the
> training data in itself is already far from pure. Honestly I'm a bit
> shocked, you would suggest I'd come up with a nostalgic argument about
> purity.
>
> Francis' examples of hip hop and dnb culture, with sampling at their
> heart, already starts to point to the problems with this statement. Culture
> has always been a project of cutting and splicing, appropriating,
> transforming, and remaking existing material. It's funny that AI
> commentators like Gary Marcus talk about GPT-3 as the 'king of pastiche'.
> Pastiche is what culture does. Indeed, we have whole genres (the romance
> novel, the murder mystery, etc) that are about reproducing certain elements
> in slightly different permutations, over and over again.
>
> Maybe it is no coincidence that I included exactly this example.
>
> Unspoken in this claim of machines 'tainting' or 'corrupting' culture is
> the idea of authenticity.
>
> I didn't claim 'tainting' or 'corrupting' culture, not even unspoken. Who
> am I to argue against the productive forces?
>
> It really reminds me of the moral panic surrounding algorithmic news and
> platform-driven disinformation, where pundits lamented the shift from truth
> to 'post-truth.'  This is not to suggest that misinformation is not an
> issue, nor that veracity doesn't matter (i.e. Rohingya and Facebook). But
> the premise of some halcyon age of truth prior to the digital needs to get
> wrecked.
>
> I agree. Only, I never equaled 'uncontaminated' to a "truth prior to the
> digital", I equaled it to a snapshot that doesn't contain material created
> by transformer models.
>
> Yes, Large language models and other AI technologies do introduce new
> conditions, generating truth claims rapidly and at scale. But rather than
> hand-wringing about 'fake news,' it's more productive to see how they
> splice together several truth theories (coherence, consensus, social
> construction, etc) into new formations.
>
> I was more interested in two points:
>
> 1.) Subversion: What I called in my original text the 'data space'
> (created through cultural snapshots as suggested by Eva Cetinic) is an
> already biased, largely uncurated information space where image data and
> language data are scaped and then mathemtically-statistically merged
> together. The focus point here is the sheer scale on which this happens.
> GPT-3 and CLIP are techniques that both build on massive datascraping
> (compared for instance to GANs) so that it is only possible for well funded
> organizations such as Open-AI or LAION to build these datasets. This
> dataspace c

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Brian Holmes]

[This was written yesterday, so it responds mostly to Luke and Felix.]

I agree that pastiche is a fundamental cultural process - but if it's so
fundamental, then to make any distinctions you have to look at its effects
in specific contexts. One such context, in the recent past, is
postmodernism. It's relevant in some ways, but I agree with Francis that
the present context is quite different.

Postmodern pastiche is the original twist that an individual gives to a
mass-distribution image. In the arts of the 1980s and 90s, the pastiche
aesthetic had the effect of disqualifying a whole range of avant-garde
practices, from neo-dadist transgression to modernist abstraction, all of
which consciously tried to mark off a space *outside* corporate-capitalist
aesthetic production. From one angle, the acceptance of a common commercial
culture was a good thing: it reduced the power of elite gatekeepers, since
the raw material of art was now ready-to-hand, without racial, financial
and educational barriers to access. But the quest for autonomy is another
fundamental cultural process, and in contemporary societies, autonomy from
highly manipulative aesthetic production is crucial. Otherwise, there's
nowhere to develop any divergent ethical/political orientation. As the
focus of commercial culture shifted online, these problems took on new
guises. Most of my own work as a cultural critic in the 2000s was devoted
to autonomy in the communication societies - and then came social media,
making the whole situation dramatically worse.

Today, Francis points to the floods of imagery that are already being
produced by AI/statistical computing, and he predicts second and third
generations of degraded images, synthesized from the initial ones. I was
struck by this word "degraded" in the initial text, and I think it
corresponds to something more than simple entropy on the level of data. The
absence of any individual or subcultural viewpoint at the origin of the
statistically generated images, and the coresponding lack of particular
affects, aspirations, insights or blindspots, renders yet another
fundamental cultural process obsolete - namely, interpretation. I question
whether the notion of pastiche makes any sense at all without
interpretation (preumably Luke has something to say about that). What's
certain is that the autonomy question is becoming urgent.

Autonomy is not about purity, nor self-sufficiency, nor withdrawal. It's
about the ability to establish the terms (particularly the overarching
value orientation) that will guide one's engagement with society. I agree
with Francis that being able to filter out statistically produced images
(and music, and discourse) is going to become a major issue under flood
conditions. And I'd go further. Whoever is not able to form or join an
interpretative community, very consciously dedicated to making meaning with
respect to art or other cultural practices, is going to experience a very
profound alienation during the next phase of the communication societies.



On Fri, Dec 23, 2022 at 8:54 AM Francis Hunger 
wrote:

> Dear Luke, dear All
>
> Interesting essay Francis, and always appreciate Brian's thoughtful
> comments. I think the historical angle Brian is pointing towards is
> important as a way to push against the claims of AI models as somehow
> entirely new or revolutionary.
>
> In particular, I want to push back against this idea that this is the last
> 'pure' cultural snapshot available to AI models, that future harvesting
> will be 'tainted' by automated content.
>
> At no point did I allude to the 'pureness' of a cultural snapshot, as you
> suggest. Why should I? I was discussing this from a material perspective,
> where data for training diffusion models becomes the statistical material
> to inform these models. This data has never been 'pure'. I used the
> distinction of uncontaminated/contaminated to show the difference between a
> training process for machine learning which builds on an snapshot, that is
> still uncontaminated by the outputs of CLIP or GPT and one which includes
> generated text and images using this techique on a large scale.
>
> It is obvious, but maybe I should have made it more clear, that the
> training data in itself is already far from pure. Honestly I'm a bit
> shocked, you would suggest I'd come up with a nostalgic argument about
> purity.
>
> Francis' examples of hip hop and dnb culture, with sampling at their
> heart, already starts to point to the problems with this statement. Culture
> has always been a project of cutting and splicing, appropriating,
> transforming, and remaking existing material. It's funny that AI
> commentators like Gary Marcus talk about GPT-3 as the 'king of pastiche'.
> Pastiche is what culture does. Indeed, we have whole genres (the romance
> novel, the murder mystery, etc) that are about reproducing certain elements
> in slightly different permutations, over and over again.
>
> Maybe it is no coincidence that I inclu

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Francis Hunger]

Dear Luke, dear All
Interesting essay Francis, and always appreciate Brian's thoughtful 
comments. I think the historical angle Brian is pointing towards is 
important as a way to push against the claims of AI models as somehow 
entirely new or revolutionary.


In particular, I want to push back against this idea that this is the 
last 'pure' cultural snapshot available to AI models, that future 
harvesting will be 'tainted' by automated content.


At no point did I allude to the 'pureness' of a cultural snapshot, as 
you suggest. Why should I? I was discussing this from a material 
perspective, where data for training diffusion models becomes the 
statistical material to inform these models. This data has never been 
'pure'. I used the distinction of uncontaminated/contaminated to show 
the difference between a training process for machine learning which 
builds on an snapshot, that is still uncontaminated by the outputs of 
CLIP or GPT and one which includes generated text and images using this 
techique on a large scale.


It is obvious, but maybe I should have made it more clear, that the 
training data in itself is already far from pure. Honestly I'm a bit 
shocked, you would suggest I'd come up with a nostalgic argument about 
purity.


Francis' examples of hip hop and dnb culture, with sampling at their 
heart, already starts to point to the problems with this statement. 
Culture has always been a project of cutting and splicing, 
appropriating, transforming, and remaking existing material. It's 
funny that AI commentators like Gary Marcus talk about GPT-3 as the 
'king of pastiche'. Pastiche is what culture does. Indeed, we have 
whole genres (the romance novel, the murder mystery, etc) that are 
about reproducing certain elements in slightly different permutations, 
over and over again.

Maybe it is no coincidence that I included exactly this example.
Unspoken in this claim of machines 'tainting' or 'corrupting' culture 
is the idea of authenticity.
I didn't claim 'tainting' or 'corrupting' culture, not even unspoken. 
Who am I to argue against the productive forces?
It really reminds me of the moral panic surrounding algorithmic news 
and platform-driven disinformation, where pundits lamented the shift 
from truth to 'post-truth.'  This is not to suggest that 
misinformation is not an issue, nor that veracity doesn't matter (i.e. 
Rohingya and Facebook). But the premise of some halcyon age of truth 
prior to the digital needs to get wrecked.
I agree. Only, I never equaled 'uncontaminated' to a "truth prior to the 
digital", I equaled it to a snapshot that doesn't contain material 
created by transformer models.
Yes, Large language models and other AI technologies do introduce new 
conditions, generating truth claims rapidly and at scale. But rather 
than hand-wringing about 'fake news,' it's more productive to see how 
they splice together several truth theories (coherence, consensus, 
social construction, etc) into new formations.


I was more interested in two points:

1.) Subversion: What I called in my original text the 'data space' 
(created through cultural snapshots as suggested by Eva Cetinic) is an 
already biased, largely uncurated information space where image data and 
language data are scaped and then mathemtically-statistically merged 
together. The focus point here is the sheer scale on which this happens. 
GPT-3 and CLIP are techniques that both build on massive datascraping 
(compared for instance to GANs) so that it is only possible for well 
funded organizations such as Open-AI or LAION to build these datasets. 
This dataspace could be spammed a) if you want to subvert it and b) if 
you'd want to advertise. The spam would need to be on a large scale in 
order to influence the next (contaminated) iteration of a cultural 
snapshot. In that sense only I used the un/contaminated distinction.


2). In response to Brian I evoked a scenario that builds on what we 
already experience when it comes to information spamming. We all know, 
that mis-information is a social and _not_ a machinic function. Maybe I 
should have made this more clear (I simply assumed it). I ignored Brians 
comment on the decline of culture, whatever this would mean, and could 
have been more precise in this regards. I don't assume culture declines. 
Beyond this, there have been discussions about deepfakes for instance 
and we saw that deepfakes are not needed at all to create 
mis-information, when one can just cut any video using standard video 
editing practices towards 'make-believe'. I wasn't 'hand-wringing' about 
fake news, in my comment to Brian, instead I was quoting Langlois with 
the concept of 'real fakes'.
Further I'm suggesting that CLIP and GPT make it more easy to automate 
large scale spamming, making online communities uninhabitable or 
moderation more difficult. Maybe I'm overestimating the effect. We can 
already observe GPT-3 automated comments appearing on twitter or the ban 
of GPTChat pos

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Felix Stalder]

I couldn't agree more. There is no such thing as authentic culture, 
particularly not on a world where desires have manufactured by consumer 
capitalism for generations.


This reminds me of a work by the Mediengruppe Bitnik, State of Reference 
(2017)


https://ww.bitnik.org/sor

It's a simple work, producing a chain of images where the last image 
serves as an input to search of a visually similar image, which is then 
becomes the input for the next search, and so on.


It's a way of navigating through the most dense nodes of Google's 
knowledge about the visual world. And it's thoroughly depressing: It 
starts from  Man Ray's The Poet (1938) only to jump immediately to stock 
images of people and products, celebrities, beauty clinics, real estate 
and some geometric figures and a few uplifting quotes. After 1321 
iterations it arrives at the image of an Nespresso machine.


None of these images as AI generated, but the commercial pollution had 
turned the image pool on which Google has trained its image recognition 
software already toxic.


I'm not suggesting that it's all the same same old, or that things 
cannot get worse,  but rather that however bad we think the current 
situation is, nostalgia is a bad form of critique.



all the best. Felix




On 21.12.22 01:26, Luke Munn wrote:
Interesting essay Francis, and always appreciate Brian's thoughtful 
comments. I think the historical angle Brian is pointing towards is 
important as a way to push against the claims of AI models as somehow 
entirely new or revolutionary.


In particular, I want to push back against this idea that this is the 
last 'pure' cultural snapshot available to AI models, that future 
harvesting will be 'tainted' by automated content.


Francis' examples of hip hop and dnb culture, with sampling at their 
heart, already starts to point to the problems with this statement. 
Culture has always been a project of cutting and splicing, 
appropriating, transforming, and remaking existing material. It's funny 
that AI commentators like Gary Marcus talk about GPT-3 as the 'king of 
pastiche'. Pastiche is what culture does. Indeed, we have whole genres 
(the romance novel, the murder mystery, etc) that are about reproducing 
certain elements in slightly different permutations, over and over again.


This is not a recent or purely digital phenomenon. I remember going to a 
show at the Neue Nationalgalerie, where oil paintings repeatedly 
reproduced the identical bird in different positions. "A variety of 
painting styles suggests the involvement of a number of assistants and 
several motifs can be repeatedly found in an unaltered form in many of 
his paintings. D’Hondecoeter’s oeuvre consequently appears as a 
conglomeration of decorative collages, produced in an almost mechanical 
seriality on the basis of successful formulas." Copy, paste, repeat.


Unspoken in this claim of machines 'tainting' or 'corrupting' culture is 
the idea of authenticity. It really reminds me of the moral panic 
surrounding algorithmic news and platform-driven disinformation, where 
pundits lamented the shift from truth to 'post-truth.'  This is not to 
suggest that misinformation is not an issue, nor that veracity doesn't 
matter (i.e. Rohingya and Facebook). But the premise of some halcyon age 
of truth prior to the digital needs to get wrecked. Yes, Large language 
models and other AI technologies do introduce new conditions, generating 
truth claims rapidly and at scale. But rather than hand-wringing about 
'fake news,' it's more productive to see how they splice together 
several truth theories (coherence, consensus, social construction, etc) 
into new formations. I'm currently writing a paper precisely on this 
issue with a couple of colleagues.


nga mihi / best,
Luke


On Tue, 20 Dec 2022 at 22:20, Francis Hunger 
mailto:francis.hun...@irmielin.org>> wrote:


Hi Brian,

On Mon, Dec 19, 2022 at 3:55 AM Francis Hunger
mailto:francis.hun...@irmielin.org>>
wrote:

While some may argue that generated text and images will save
time and money for businesses, a data ecological view
immediately recognizes a major problem: AI feeds into AI. To
rephrase it: statistical computing feeds into statistical
computing. In using these models and publishing the results
online we are beginning to create a loop of prompts and
results, with the results being fed into the next iteration of
the cultural snapshots. That’s why I call the early cultural
snapshots still uncontaminated, and I expect the next
iterations of cultural snapshots will be contaminated.


Francis, thanks for your work, it's always totally interesting.

Your argumentation is impeccable and one can easily see how
positive feedback loops will form around elements of AI-generated
(or perhaps "recombined") images. I agree, this will become
untenable, though I'd be interested in

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Luke Munn]

Interesting essay Francis, and always appreciate Brian's thoughtful
comments. I think the historical angle Brian is pointing towards is
important as a way to push against the claims of AI models as somehow
entirely new or revolutionary.

In particular, I want to push back against this idea that this is the last
'pure' cultural snapshot available to AI models, that future harvesting
will be 'tainted' by automated content.

Francis' examples of hip hop and dnb culture, with sampling at their heart,
already starts to point to the problems with this statement. Culture has
always been a project of cutting and splicing, appropriating, transforming,
and remaking existing material. It's funny that AI commentators like Gary
Marcus talk about GPT-3 as the 'king of pastiche'. Pastiche is what culture
does. Indeed, we have whole genres (the romance novel, the murder mystery,
etc) that are about reproducing certain elements in slightly different
permutations, over and over again.

This is not a recent or purely digital phenomenon. I remember going to a
show at the Neue Nationalgalerie, where oil paintings repeatedly reproduced
the identical bird in different positions. "A variety of painting styles
suggests the involvement of a number of assistants and several motifs can
be repeatedly found in an unaltered form in many of his paintings.
D’Hondecoeter’s oeuvre consequently appears as a conglomeration of
decorative collages, produced in an almost mechanical seriality on the
basis of successful formulas." Copy, paste, repeat.

Unspoken in this claim of machines 'tainting' or 'corrupting' culture is
the idea of authenticity. It really reminds me of the moral panic
surrounding algorithmic news and platform-driven disinformation, where
pundits lamented the shift from truth to 'post-truth.'  This is not to
suggest that misinformation is not an issue, nor that veracity doesn't
matter (i.e. Rohingya and Facebook). But the premise of some halcyon age of
truth prior to the digital needs to get wrecked. Yes, Large language models
and other AI technologies do introduce new conditions, generating truth
claims rapidly and at scale. But rather than hand-wringing about 'fake
news,' it's more productive to see how they splice together several truth
theories (coherence, consensus, social construction, etc) into new
formations. I'm currently writing a paper precisely on this issue with a
couple of colleagues.

nga mihi / best,
Luke


On Tue, 20 Dec 2022 at 22:20, Francis Hunger 
wrote:

> Hi Brian,
>
> On Mon, Dec 19, 2022 at 3:55 AM Francis Hunger <
> francis.hun...@irmielin.org> wrote:
>
>> While some may argue that generated text and images will save time and
>> money for businesses, a data ecological view immediately recognizes a major
>> problem: AI feeds into AI. To rephrase it: statistical computing feeds into
>> statistical computing. In using these models and publishing the results
>> online we are beginning to create a loop of prompts and results, with the
>> results being fed into the next iteration of the cultural snapshots. That’s
>> why I call the early cultural snapshots still uncontaminated, and I expect
>> the next iterations of cultural snapshots will be contaminated.
>>
>
> Francis, thanks for your work, it's always totally interesting.
>
> Your argumentation is impeccable and one can easily see how positive
> feedback loops will form around elements of AI-generated (or perhaps
> "recombined") images. I agree, this will become untenable, though I'd be
> interested in your ideas as to why. What kind of effects do you foresee,
> both on the level of the images themselves and their reception?
>
> Foresight is a difficult field, as most estimates can extrapolate maximum
> 7 year into the future and there are a lot of independent factors (such as
> e.g. OpenAI, the producer of CLIP could go bankrupt etc.).
>
> It's worth considering that similar loops have been in place for decades,
> in the area of market research, product design and advertising. Now, all of
> neoclassical economics is based on the concept of "consumer preferences,"
> and discovering what consumers prefer is the official justification for
> market research; but it's clear that advertising has attempted, and in many
> cases succeeded, in shaping those preferences over generations. The
> preferences that people express today are, at least in part, artifacts of
> past advertising campaigns. Product design in the present reflects the
> influence of earlier products and associated advertising.
>
> That's an great and interesting argument. Because it plays into the
> cultural snapshot idea.
>
> Obviously Language wise, people already use translation tools, such as
> Deepl and translate Text from German to English and back to German in order
> to profit off the "clarity" and "orthographic correction" brought by the
> statistical analysis that feeds into the translator and seems to straighten
> the German text. We see the same stuff appearing for products like text

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Francis Hunger]

Hi Brian,
On Mon, Dec 19, 2022 at 3:55 AM Francis Hunger 
 wrote:


While some may argue that generated text and images will save time
and money for businesses, a data ecological view immediately
recognizes a major problem: AI feeds into AI. To rephrase it:
statistical computing feeds into statistical computing. In using
these models and publishing the results online we are beginning to
create a loop of prompts and results, with the results being fed
into the next iteration of the cultural snapshots. That’s why I
call the early cultural snapshots still uncontaminated, and I
expect the next iterations of cultural snapshots will be contaminated.


Francis, thanks for your work, it's always totally interesting.

Your argumentation is impeccable and one can easily see how positive 
feedback loops will form around elements of AI-generated (or perhaps 
"recombined") images. I agree, this will become untenable, though I'd 
be interested in your ideas as to why. What kind of effects do you 
foresee, both on the level of the images themselves and their reception?


Foresight is a difficult field, as most estimates can extrapolate 
maximum 7 year into the future and there are a lot of independent 
factors (such as e.g. OpenAI, the producer of CLIP could go bankrupt etc.).


It's worth considering that similar loops have been in place for 
decades, in the area of market research, product design and 
advertising. Now, all of neoclassical economics is based on the 
concept of "consumer preferences," and discovering what consumers 
prefer is the official justification for market research; but it's 
clear that advertising has attempted, and in many cases succeeded, in 
shaping those preferences over generations. The preferences that 
people express today are, at least in part, artifacts of past 
advertising campaigns. Product design in the present reflects the 
influence of earlier products and associated advertising.


That's an great and interesting argument. Because it plays into the 
cultural snapshot idea.


Obviously Language wise, people already use translation tools, such as 
Deepl and translate Text from German to English and back to German in 
order to profit off the "clarity" and "orthographic correction" brought 
by the statistical analysis that feeds into the translator and seems to 
straighten the German text. We see the same stuff appearing for products 
like text editors and thus widely employed for cultural production. 
That's one example. Automated forum posts using GPT-3, for instance on 
Reddit are another, because we know that the CLIP Model also partly 
build on Reddit posts.


Another example is images generated using diffusion models and prompts 
building on cultural snapshots and being used as _cheap_ illustrations 
for editorial products, feeding off stock photography and to a certain 
extend replacing stock photography. This is more or less an economic 
motivation with cultural consequences. The question is what changes, 
when there is not sufficiently 'original' stock photography circulating, 
but the majority is syntheticly generated? Maybe others want to join in, 
to speculate about it.


We could further look into 1980s HipHop or 1990s Drum'n Bass sample 
culture, which for instance took (and some argue: stole) one particular 
sound break, the Amen Break, from an obscure 1969 Soul music record by 
The Winston Brothers and build a whole cultural genre from it. Cf. 
https://en.wikipedia.org/wiki/Amen_break Here the sample was refined 
over time, with generations of musicians cleaning the sample 
(compression, frequencies, deverbing, etc.) and providing many 
variations of it, then reusing it, because later generation did not 
build on the original sample, but on the published versions of it.


We can maybe distinguish two modi operandi where a) "the cultural 
snapshot" is understood as an automated feedback loop, operating on a 
large scale, mainly through automated scraping and publication of the 
derivates of data, amplifying the already most visible representations 
of culture and b) "the cultural snapshot" is a feedback loop with many 
creative human interventions, be it through curatorial selection, prompt 
engineering or intended data manipulation.


Blade Runner vividly demonstrated this cultural condition in the early 
1980s, through the figure of the replicants with their implanted 
memories.
I dont know if I get your point. I'd always say that Blade Runner is a 
cultural imaginary, one of the many phantasms about the machinisation of 
humans since at least 1900 if not earlier, and that's an entirely 
different discussion then. I would avoid this as an metaphor.
The intensely targeted production of postmodern culture ensued, and 
has been carried on since then with the increasingly granular market 
research of surveillance capitalism, where the calculation of 
statistically probable behavior becomes a good deal more precise. The 
effect across the

Re: Spamming the Data Space – CLIP, GPT and synthetic data

[Brian Holmes]

On Mon, Dec 19, 2022 at 3:55 AM Francis Hunger 
wrote:

> While some may argue that generated text and images will save time and
> money for businesses, a data ecological view immediately recognizes a major
> problem: AI feeds into AI. To rephrase it: statistical computing feeds into
> statistical computing. In using these models and publishing the results
> online we are beginning to create a loop of prompts and results, with the
> results being fed into the next iteration of the cultural snapshots. That’s
> why I call the early cultural snapshots still uncontaminated, and I expect
> the next iterations of cultural snapshots will be contaminated.
>

Francis, thanks for your work, it's always totally interesting.

Your argumentation is impeccable and one can easily see how positive
feedback loops will form around elements of AI-generated (or perhaps
"recombined") images. I agree, this will become untenable, though I'd be
interested in your ideas as to why. What kind of effects do you foresee,
both on the level of the images themselves and their reception?

It's worth considering that similar loops have been in place for decades,
in the area of market research, product design and advertising. Now, all of
neoclassical economics is based on the concept of "consumer preferences,"
and discovering what consumers prefer is the official justification for
market research; but it's clear that advertising has attempted, and in many
cases succeeded, in shaping those preferences over generations. The
preferences that people express today are, at least in part, artifacts of
past advertising campaigns. Product design in the present reflects the
influence of earlier products and associated advertising.

One of the primary fields of production in the overdeveloped societies is
the field or product range of culture itself, such as movies and TV shows.
In the case of TV, feedback loops have been employed systematically since
the early 1950s, with the introduction of Nielsen's audiometer, a device
that was directly attached to thousands of TVs. Today, TV shows and
especially movies are not only used to define the cultural context of
successive "generations'' (Gen X, etc). Marketers also use them as
surrogates for the memories and affects of those generations. Of course
these proxy memories cannot cover the full range of generational
experience, but they have the immense advantage, for advertisers, of being
fully knowable and therefore, calculable in their effects. The calculations
may be more or less bullshit, but they are still employed and acted upon.

Blade Runner vividly demonstrated this cultural condition in the early
1980s, through the figure of the replicants with their implanted memories.
The intensely targeted production of postmodern culture ensued, and has
been carried on since then with the increasingly granular market research
of surveillance capitalism, where the calculation of statistically probable
behavior becomes a good deal more precise. The effect across the neoliberal
period has been, not increasing standardization or authoritarian control,
but instead, the rationalized proliferation of customizable products, whose
patterns of use and modification, however divergent or "deviant" they may
be, are then fed back into the design process. Not only the "quality of the
image" seems to degrade in this process. Instead, culture in general seems
to degrade, even though it also becomes more inclusive and more diverse at
the same time.

AI is poised to do a lot of things - but one of them is to further
accelerate the continual remaking of generational preferences for the needs
of capitalist marketing. Do you think that's right, Francis? What other
consequences do you see? And above all, what to do in the face of a
seemingly inevitable trend?

best, Brian
#  distributed via : no commercial use without permission
#is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: http://mx.kein.org/mailman/listinfo/nettime-l
#  archive: http://www.nettime.org contact: nett...@kein.org
#  @nettime_bot tweets mail w/ sender unless #ANON is in Subject:

Spamming the Data Space – CLIP, GPT and synthetic data

[Francis Hunger]

Dear Nettimers,

honoring the institutionalized format, I'm posting this speculative text 
in the hope for comments.


best

Francis

@databasecultures@dair-community.social / www.irmielin.org



*** Spamming the Data Space – CLIP, GPT and synthetic data ***
** Introduction **

For the last time in human history the cultural-data space has not been 
contaminated. In recent years a new technique to acquire knowledge has 
emerged. Scraping the Internet and extracting information and data has 
become a new modus for companies and for university researchers in the 
field of machine learning. One of the currently largest publicly 
available training data sets to combine images and labels (which shall 
describe the images content), is Laion-5B, with 5,85 billion image-text 
pairs (Ilharco, Gabriel et al. 2021).[1]
The scope of scraping internet resources has become so all-encompassing, 
that researcher Eva Cetinic has proposed to call this form ‘cultural 
snapshot’: “By encoding numerous associations which exist between data 
items collected at a certain point in time, those models therefore 
represent synchronic assemblages of cultural snapshots, embedded in a 
specific technological framework. Metaphorically those models can be 
considered as some sort of encapsulation of the collective (un)conscious 
[…]” (Cetinic 2022).[2] The important suggestion which Cetinic makes, is 
that these data collections are temporally anchored. The temporal 
dimension of these snapshots suggests that digital cultural snapshots 
taken at different times document different states of (online-)culture. 
So how will a 2021 snapshot differ from a 2031 cultural snapshot?


** Consequences **

Multi-modal models, like CLIP, trained on large-scale data sets, such as 
LAION-5B provide the statistical means to generate images from text 
prompts. In the CLIP Model, pre-trained models merge two embedding 
spaces, one for images and one for text-descriptions which with 
mathematical methods get layered together, so that the vectors in the 
one space, the image domain, align with vectors in the other space, the 
text domain, assuming there is a similarity between both, and one can 
translate into the other. In three short examples I’ll discuss some of 
the consequences of the underlying data for large-scale models from the 
perspective of cultural snapshots.


1.) Data Bias: A critical discussion of these large-scale multi-modal 
models for instance, has pointed out how they are culturally skewed and 
reproduce sexist and racist biases. Researchers Fabian Offert and Thao 
Phan, for instance, describe how the company Open AI decided not to 
mitigate the problem of whiteness by changing the model’s underlying 
data. Instead, Open AI added certain invisible keywords to users’ 
prompts to have more people of color included, without changing the 
model. Obviously, the calculations for creating these models or even 
curating the underlying data are so tremendous that for economic reasons 
even major problems cannot be corrected in the embedding space itself. 
Discussing the prevalent ‘whiteness’ in these models further, Offert and 
Phan suggest to turn to humanities in order to “identify the different 
technical modes of whiteness at play, and understand the 
reconceptualization and resurrection of whiteness as a machinic concept” 
(Offert and Phan 2022, 3).[3]


2.) Uneven spatial distribution: Users of large-scale multi-modal models 
have tested their limits when generating images. ‘Crungus’, and ‘Loab’ 
are two examples. ‘Loab’, the image of a women appeared when AI artist 
Supercomposite looked for the negative of a prompt: “DIGITA PNTICS 
skyline logo::-1”. Loab appears to be a consistent pixel accumulation, 
which repeatedly emerges in different configurations and cannot easily 
be traced back to a single origin.[4] The creator/discoverer of ‘Loab’ 
felt during intensive testing, that Loab might exist in its own pocket, 
because it was relatively reproducible, compared to other prompts, as if 
it was populating a certain statistical region within the larger latent 
space. Another, similar phenomenon of uneven spatial distribution in 
latent space is ‘Crungus’, basically a phantasy word which as a prompt 
nevertheless created results: a snarling, zombie-like figure with 
shoulder-long hair, which could be part of a horror movie.[5]


Both examples demonstrate that the cultural snapshots also contain 
material which cannot be easily identified or traced back and they 
demonstrate, how the latent space is an uneven spatial distribution by 
design. Since the models are built by a process called zero shot 
learning in difference to for instance the supervised learning used in 
ImageNet, there are no longer intentional ontologies used in the 
knowledge creation of these models. The human involvement involves the 
uncoordinated captioning of images by users online, and the setting up 
the scraping algorithms and excluding certain domains from