Series: Hyperreality Page 4 of 20

Examining the blending of fact & fiction in online, augmented, and hyperreal environments.

Disorientation in AI writing

On 14 September 2022

There’s a really good article on the Verge about authors who use AI tools like Sudowrite as part of their writing workflow. Lost Books has released about a dozen books in this genre now, which comprise the AI Lore series.

Anyway, there are a few themes I want to tease out, namely the feeling of disconnection & disorientation that seems to be a common experience to authors using these tools.

One author quoted says:

“It was very uncomfortable to look back over what I wrote and not really feel connected to the words or the ideas.”

And:

“But ask GPT-3 to write an essay, and it will produce a repetitive series of sometimes correct, often contradictory assertions, drifting progressively off-topic until it hits its memory limit and forgets where it started completely.”

And finally:

And then I went back to write and sat down, and I would forget why people were doing things. Or I’d have to look up what somebody said because I lost the thread of truth,” she said.

Losing the “thread of truth” strikes me as utterly & inherently postmodern af. It’s the essence of hyperreality.

It is the essence of browsing the web. You pop between tabs and websites and apps and platforms. You follow different accounts, each spewing out some segment of something. And then somewhere in the mix, your brain mashes it all together into something that sort of makes sense to you in its context (“sensemaking”), or doesn’t — you lose the thread of truth.

To me, hyperreality as an “art form” (way of life?), has something to do with that. To the post-truth as they say world where truth is what resonates in the moment. What you “like” in platform speak, what you hate, what you fear, just then, just now. And then its forgotten, replaced by the next thing. Yet the algorithm remembers… or does it? It may be “recorded” but it knows little to nothing on its own, without the invocation.

Forgive me as I ramble here, but that’s why this is a blog post…

Pieces I’ve been meaning to put together in this space.

Narrative unity
Networked / distributed narratives
Transmedia storytelling
Latent space (as in Stable Diffusion)
Jacques Vallee’s idea of the information-based universe

In no particular order:

“Networked narratives can be seen as being defined by their rejection of narrative unity.”
https://en.wikipedia.org/wiki/Networked_narrative

The PDF wikipedia goes on to reference regarding narrative unities has some worthwhile stuff on the topic. From it, we see these are more properly perhaps called Dramatic Unities (via Aristotle — an ancient blogger if ever there was one), or Wiki’s redirect to Classical Unities here.

1. unity of action: a tragedy should have one principal action.

2. unity of time: the action in a tragedy should occur over a period of no more than 24 hours.

3. unity of place: a tragedy should exist in a single physical location.

Popping back to the W networked narrative page:

“It is not driven by the specificity of details; rather, details emerge through a co-construction of the ultimate story by the various participants or elements.”

Lost the thread of truth, “not driven by the specificity of details.”

While we’re in this uncertain territory, we should at least quote again Wikipedia on Hyperreality:

“Hyperreality is seen as a condition in which what is real and what is fiction are seamlessly blended together so that there is no clear distinction between where one ends and the other begins.”

What I guess I want to note here – in part – is that what the Verge article quoted at top seems to be considering obstacles, bugs, or room for improvement around the lack of apparent coherence of AI-generated texts… is actually probably its primary feature?

Disorientation as a Service.

Jumping now to latent spaces, as in AI image generation:

“A latent space, also known as a latent feature space or embedding space, is an embedding of a set of items within a manifold in which items which resemble each other more closely are positioned closer to one another in the latent space.”

This Vox video is probably the most complete and accessible explanation I’ve seen of how image diffusion models work:

My understanding of it is basically that a text query (in the case of Dall-E & Stable Diffusion) triggers accessing of the portion(s) of the latent space within the model that corresponds to your keywords, and then mashes them together visually to create a cloud of pixels that reference those underlying trained assets. Depending on your level of processing (“steps” in Stable Diffusion), the diffuse pixel cloud becomes more precisely representative of some new vignette that references your original query or prompt.

So it sort of plucks what you asked for out of its matrix of possible combinations, and gives you a few variations of it. Kind of like parallel dimension representations from the multiverse.

Which leads me to the Jacques Vallee quote that has been stirring around in the corners of my mind for some twenty-odd years now:

Time and space may be convenient notions for plotting the progress of a locomotive, but they are completely useless for locating information …

What modern computer scientists have now recognized is that ordering by time and space is the worst possible way to store data. In a large computer-based information system, no attempt is made to place related records in sequential physical locations. It is much more convenient to sprinkle the records through storage as they arrive, and to construct an algorithm for the retrieval based on some kind of keyword …

(So) if there is no time dimension as we usually assume there is, we may be traversing events by association.

Modern computers retrieve information associatively. You “evoke” the desired records by using keywords, words of power: (using a search engine,) you request the intersection of “microwave” and “headache,” and you find twenty articles you never suspected existed … If we live in the associative universe of the software scientist rather than the sequential universe of the spacetime physicist, then miracles are no longer irrational events.

Vallee’s quote strikes a metaphysical chord which is mostly unprovable (for now) but also feels , experientially speaking, “mostly true” in some ways. Without debating the ontological merits of his argument vis-a-vis everyday reality, it occurs to me that he’s 100% describing the querying of latent spaces.

Of course, he suggests that reality consists of a fundamental underlying latent space, which is a cool idea if nothing else. There’s an interesting potential tangent here regarding paranormal events and “retrieval algorithms” as being guided by or inclusive of intelligences, perhaps artificial, perhaps natural. (And that tangent would link us back to Rupert Sheldrake’s morphogenetic/morphic fields as retrieval algorithms, and maybe the “overlighting intelligences” of Findhord…) But that’s a tangent for another day.

Anyway, to offer some sort of conclusion, I guess I would say perhaps the best use of AI tools for now, while they are in their current form, is to lean into, chase after, capture that confusion, that disorientation, that losing of the thread, that breaking of narrative unity, and just… go for it. There are as many roads through the Dark Forest as we make.

The Dissolution of Meaning

By Tim B.

On 2 November 2022

In Other

A lot of times, I will search Google for something, click through to a page that “seems” like information, and then discover in a surface skim that it’s actually basically junk and/or trying to sell you a product above and beyond the mere SEO manipulation. In those cases, I feel had to a certain extent–even if the failure is necessarily in many way’s Google’s for bringing me this junk in the first place and trying to hide it among or in place of “real” meaning and information.

Which of course pushes my heavy experimentation with AI writing tools to produce books into a certain state of tension. I know that; I own it. It’s the uncanny valley of delight and terror that I choose to play in. Because I know in that tension itself is something to be unwound and explored.

If a book is wholly or partially written by an AI, what impact does that actually have on it? Is it “better” or “worse” in some way, because there is either a lesser or else different impulse behind it’s creation? Is it more or less “worthwhile” or “valuable?”

In my case (and I should preface this by saying that I don’t necessarily consider myself or strictly care about “authorship”–that’s a hang up I’ve chosen to put aside…), I see these books as an interrogation of the technologies themselves. Personally I don’t like when people call the tools in their current state “AIs.” I feel that’s a tremendous overshoot when it’s really just machine learning applied at various scales. But that’s a subtlety that’s lost on the masses who just want a good headline to click on, and then ignore the article’s actual contents.

Which is a pattern we’re all used to. It’s, in a way, fundamental I think to the hyperlink, though it had to be laundered through a decade or two of dirtying human nature first to become really readily apparent.

I don’t really agree with that one dude’s estimation that Lambda is a “sentient” chat bot, but I’ve played with others enough to know that there is a spooky effect here, probably latent in human consciousness, or in material-cosmic consciousness itself. We’re gonna project our own meaning into it, even if that meaning is “this is crap,” or “this is fake”;–all valid reactions. Just as much as this is fun or this is good.

Why shouldn’t we ask these technologies, though, what they “think,” leaving aside their actual ontological status (which is unknowable)? And just see what they say, and then ask them more questions, and more.

What if the answers they give are “wrong” or “false” or “bad” or “dangerous?” What if they are misinformation or “disinformation”, or advocate criminal acts, or suicide?

The problem is the dissolution of meaning, to which these are only an accelerant, not the underlying cause (though they will certainly fuel a feedback loop). These tools are terrible at holding a narrative thread, of keeping track of characters in a scene, what’s going on, or how we got here, let alone where we are going. In a way, that’s freeing, to smash narrative unity. I don’t think I’m the first creator to discover this freedom, either.

Wikipedia:

“Surrealism is a cultural movement that developed in Europe in the aftermath of World War I in which artists depicted unnerving, illogical scenes and developed techniques to allow the unconscious mind to express itself.^[1] Its aim was, according to leader André Breton, to “resolve the previously contradictory conditions of dream and reality into an absolute reality, a super-reality”, or surreality.“

And surrealism’s cousin Dadaism:

“Developed in reaction to World War I, the Dada movement consisted of artists who rejected the logic, reason, and aestheticism of modern capitalist society, instead expressing nonsense, irrationality, and anti-bourgeois protest in their works.”

Moving on to hyperrealism (and I will trot this quote out endlessly for ever and ever, amen):

“Hyperreality is seen as a condition in which what is real and what is fiction are seamlessly blended together so that there is no clear distinction between where one ends and the other begins.”

I like this idea that there is a reality above reality in which contradictions are merged, and taken rightly as significant elements of but a greater whole which encompasses all elements, much like the Hypogeum of Quatrian lore. That what is “real” and “unreal” are merely glimpses on a continuum of experience itself.

If there is any beauty or truth to be found in any of those arts, then there must be too in the merging and dissolution of meaning and non-meaning that we see so strongly emergent as a current in AI-produced art (including literature).

Why not let the reader’s confusion about what was written by AI and what by human be a part of the longing they have to develop and nurture, that desire to understand, or at least swim or float in the sea of non-meaning?

Does it degrade meaning? Does it uplift non-meaning? Non-meaning is another form of meaning. “Alternative facts,” alternative fictions. Who picks up and leaves off? You? A robot? A corporation? A government? Each an authority, each offering their assessment. All of which one could take or leave, depending on who has the arms and means of enforcement. The game then begins I guess to be simply how to navigate these waters, how not to get too hung, how not to get too exploited, how not to be too weighed down when you dive into the depths of rabbit holes under the sea that you can come back up again, and still be free.

Free to find bad search results. Free to be mislead, and freedom to mislead. All is sleight of hand and stage direction. Everything gardens, trying to manipulate its environment to create optimal conditions for its own survival. Including AI, including all of us. We need new tools to understand. We need brave explorers to sail off into unmeaning, and bring back the treasures there, for not all life is laden and wonderful. Much is viral and stupid. Much is lost, much to be gained.

On “Dangerous” fictions

By Tim B.

On 28 December 2022

In Other

Found this piece from July 2022 by Cory Doctorow, where he talks about an author who was apparently a protege of Philip K. Dick’s who I never heard of – Tim Powers.

In it, he brings up an oft-repeated trope regarding “dangerous” fictions, a pet topic of mine:

“The Powers method is the conspiracist’s method. The difference is, Powers knows he’s making it up, and doesn’t pretend otherwise when he presents it to us. […]

The difference between the Powers method and Qanon, then, is knowing when you’re making stuff up and not getting high on your own supply. Powers certainly knows the difference, which is why he’s a literary treasure and a creative genius and not one of history’s great monsters.”

As popular as this type of argument is (and Douglas Rushkoff trots out something similar here and here), I personally find it to be overly simplistic and a bit passé.

First of all, I would argue that all writers – by necessity – must get “high on their own supply” in order to create (semi) coherent imaginal worlds and bring them to fruition for others to enjoy. Looking sternly at you here, Tolkien. In fact, perhaps the writers who get highest on their own supply are in some cases the best…

Second, no one arguing in favor of this all of nothing position (fiction must be fiction must be fiction) seems to have taken into account the unreliable narrator phenomenon in fiction.

Wikipedia calls it a narrator whose credibility is compromised:

“Sometimes the narrator’s unreliability is made immediately evident. For instance, a story may open with the narrator making a plainly false or delusional claim or admitting to being severely mentally ill, or the story itself may have a frame in which the narrator appears as a character, with clues to the character’s unreliability. A more dramatic use of the device delays the revelation until near the story’s end. In some cases, the reader discovers that in the foregoing narrative, the narrator had concealed or greatly misrepresented vital pieces of information. Such a twist ending forces readers to reconsider their point of view and experience of the story. In some cases the narrator’s unreliability is never fully revealed but only hinted at, leaving readers to wonder how much the narrator should be trusted and how the story should be interpreted.”

My point is that the un/reliability of the “narrator” can extend all the way out through to the writer themself. (And what if the reader turns out to be unreliable?)

Can we ever really know for certain if a writer “believed” that thing x that they wrote was wholly fictional, wholly non-fictional, or some weird blend of the two? Do we need to ask writers to make a map of which elements of a story are which? Isn’t that in some sense giving them more power than they deserve?

Moreover, if the author is an unreliable narrator (and to some extent every subjective human viewpoint is always an unreliable narrator to some degree), how can we ever trust them to disclose to us responsibly whether or not they are indeed unreliable? Short answer is: we can’t. Not really.

This is one of those “turtles all the way down” arguments, in which (absent other compelling secondary evidence) it may be difficult or sometimes impossible to strike ground truth.

All of this boils down for me to the underlying argument of whether one must label fictional works as fiction, and if not doing so is somehow “dangerous.”

The Onion’s Amicus Brief earlier this year why parody and satire should not be required to be overtly labelled – because if robs these millennia-old art forms of their structural efficacy, their punch as it were.

Wikipedia’s Fiction entry’s history section is sadly quite scant about the details. A couple of other sources point to more specifically the 12th century in Europe (though likely it goes back farther). One source whose credibility I have no concept of states:

“In the Middle Ages, books were perceived as exclusive and authoritative. People automatically assumed that whatever was written in a book had to be true,” says Professor Lars Boje…

It’s an interesting idea, that structurally the phenomenon of the book was so rare and complex that by virtue of its existence alone, it was conceived of as containing truth.

“

Up until the High Middle Ages in the 12th century, books were surrounded by grave seriousness.

The average person only ever saw books in church, where the priest read from the Bible. Because of this, the written word was generally associated with truth.”

That article alludes to an invisible “fiction contract” between writer and reader, which didn’t emerge as a defined genre distinction until perhaps the 19th century. They do posit a transition point through in the 12th, but don’t back it up by any evidence therein of a “fiction contract.”

“The first straightforward work of fiction was written in the 1170s by the Frenchman Chrétien de Troyes. The book, a story about King Arthur and the Knights of the Round Table, became immensely popular.”

HistoryToday.com – another site whose credibility I cannot account for – seems to agree with pinpointing that genre of Arthurian romance as being linked to the rise of fiction, though pushes it back a few years to 1155, with Wace’s translation of Monmouth’s History of the Kings of Britain. The whole piece is an excellent read, so I won’t rehash it here, but quote:

“This is the literary paradigm which gives us the novel: access to the unknowable inner lives of others, moving through a world in which their interior experience is as significant as their exterior action.”

They suggest that fiction – in some form like we might recognize it today – had precursor conditions culturally that had to be met before it could arise, namely that the inner lives of people mattered as much as their outward action.

“It need hardly be said that the society which believes such things, which accedes to – and celebrates – the notion that the inner lives of others are a matter of significance, is a profoundly different society from one that does not. There is an immediately ethical dimension to these developments: once literature is engaged in the (necessarily fictional) representation of interior, individuated selves, who interact with other interior, individuated selves, then moral agency appears in a new light. It is only in the extension of narrative into the unknowable – the minds of others – that a culture engages with the moral responsibility of one individual toward another, rather than with each individual’s separate (and identical) responsibilities to God, or to a king.”

It’s interesting also here to note that, A) the King Arthur stories did not originate with Chretien de Troyes or Geoffrey of Monmouth, and B) many people ever since still believe them to be true today to some extent.

Leaving that all aside, one might also ask regarding my own work, well isn’t this all just a convoluted apologia for the type of writing I’m doing? Absolutely, and why not articulate my purpose. You can choose to believe me or decide that I am an unreliable narrator. It’s up to you. I respect your agency, but I also want to play on both the reader’s and the author’s (myself) expectations about genres and categories. These are books which take place squarely in the hyperreal after all, the Uncanny Valley. They intentionally invite these questions, ask you to suspend your disbelief, and then cunningly deconstruct it, only to reconstruct it and smash it again later – and only if you’re listening.

Further, as artists I believe our role and purpose is to some extent to befuddle convention, and ask questions that have no easy answers. Yes, this will cause some uneasiness, especially among those accustomed to putting everything into little boxes, whose contents never bleed or across. Some people might even worry if it’s “dangerous” to believe in things that aren’t factual. Is it? I think the answer is sometimes, and it depends. But it largely depends on your agency as the reader, and what you do with it in real life.

Consider the case of this purveyor of tall tales, Randy Cramer, who claims with a straight face to have spent 17 years on the Planet Mars fighting alien threats to Earth.

He is the very definition of the unreliable narrator, whose labels of fact of fiction likely do not accord with consensus reality on many major points.

The video below is a good, if a bit annoying, take-down of many of Cramer’s claims, though unfortunately I think leans rather too heavily on deconstructing his body language, when his words alone are damning enough (btw, looks like the George Noory footage comes from an interview he did for his show Beyond Belief):

The question remains: is this an example of a “dangerous” fiction?

To understand that, I tend to think in terms of risk analysis, in which we might try to estimate:

The specific harm(s)
Their likelihood of occurring
Their severity

One definition of harm traces back to Feinberg, and is something like wrongful setbacks of interest. A Stanford philosophy site further elucidates, quoting Feinberg:

Feinberg’s defines harm as “those states of set-back interest that are the consequence of wrongful acts or omissions by others” (Feinberg 1984)

Is saying you spent 17 years on Mars a “wrongful act or omission?” Perhaps. But as the Stanford article points out, actually defining what is or isn’t in someone’s interests is incredibly squishy.

In Cramer’s case, perhaps it is willfully and wrongfully deceptive to say the things he is saying. Do we have a moral or legal responsibility to always tell the truth? What about when that prevarication leads to financial loss in others?

In Cramer’s case, according to the second video linked above, he does seem to ask people for money – both in funding creation of a supposedly holographic bio-medical bed which can regrow limbs, and in the form of online psionics courses and one-on-one consultations.

But is it wrongful if the buyers/donators have agency, and the ability to reasonably evaluate his claims on their own?

Wikipedia’s common-language definition of fraud seems like it could apply here:

“…fraud is intentional deception to secure unfair or unlawful gain, or to deprive a victim of a legal right.”

Is Cramer a fraud? Is he a liar? I wondered here if Cramer might have a defamation case against the YouTube author referenced above, who calls him a pathological liar. But last time I checked, truth is an absolute defense against defamation claims. That is, the commonly accepted truth we agree on as a society – more or less – is that Mars is uninhabited, and there is no Secret Space program, etc. So if it went to court, it seems like the defamation claim would not have a leg to stand on.

Of course, it’s *possible* it’s all truth, and what we call consensus reality is based on a massive set of lies itself that is very different from ‘actual’ reality. But that’s not how courts work.

What if Cramer included disclaimers like you might see on tarot card boxes, or other similar novelty items, “For entertainment purposes only?” It depends what authority we’re trying to appeal to here: a court of law, the court of public opinion, or one reader’s experience of a particular work. Each of those might see the matter in a different light, depending on their viewpoint.

In my case, I include disclaimers regarding the inclusion of AI generated elements. I leave it up to the reader to try to determine A) which parts, and B) what the implications of AI content even are. Should they be trusted?

My position, and the one which I espouse throughout, is that – for now – AI is an unreliable narrator. Making it about on par with human authors in that regard. Are the fictions it produces “dangerous?” Must we label them “fictions” and point a damning finger at their non-human source?

In some ways, my books are both an indictment of and celebration of AI authorial tools, and even full-on AI authorship (which I think we’re some ways away from still). To know their dangers, we must probe them, and expose them thoughtfully. We must see them as they are – as both authors and readers – warts and all. And decide what we will do with the risks and harms they may pose, and how we can balance all that with an enduring belief and valorisation of human agency.

Because if we can’t trust people to make up their own minds about things they read, we run the real risk of one of the biggest and most dangerous fictions of all – that we would be better off relying on someone else to tell us what’s ‘safe’ and therefore good, and trust them implicitly to keep away anything deemed ‘dangerous’ by the authority in whom we have invested this awesome power.

Authorless writing

By Tim B.

On 30 January 2023

In Other

Something I’ve seen working in the “disinformation industrial complex” is that people after years of this proliferating online are still grappling with basic typology around the three allied terms of disinformation, misinformation, and malinformation.

A Government of Canada Cybersecurity website offers sidebar definitions of the three, clipped for brevity here:

Misinformation: “false information that is not intended to cause harm…”
Disinformation: “false information that is intended to manipulate, cause damage…” [etc]
Malinformation: “information that stems from the truth but is often exaggerated in a way that misleads…”

The two axes these kinds of analyses tend to fall on are truthfulness and intent. Secondary to that is usually harm as a third axis, which ranges from potential to actual.

Having spent a lot of time doing OSINT and content moderation work, it is very common in the field that an analyst cannot make an authoritative claim to have uncovered the absolute “truth” of something. Sometimes facts are facts, but much of the time, they become squishy “facts” which may have greater or lesser degrees of trustworthiness, depending on one’s perspective, and how much supporting data one has amassed, and the context in which they are used.

Even more difficult to ascertain in many/most cases is intent. There are so many ways to obscure or disguise one’s identity online; invented sources may be built up over years and years to achieve a specific goal, taking on the sheep’s clothing of whatever group they are trying to wolf their way into. Intent is extremely opaque, and if you do find “evidence” of it in the world of disinformation, it is very likely that it is manufactured from top to bottom. Or not, it could just be chaotic, random, satire, etc. Or just someone being an idiot and spouting off on Facebook.

Having butted up against this issue many times, I’ve switched wholly over to the “intends to or does” camp of things. Whether or not author x intended outcome y, it is observable that a given effect is happening. Then you can start to make risk assessments around the actual or probable harms, who is or might be impacted, and the likelihood and severity of the undesirable outcomes.

It’s a much subtler and more complex style of analysis, but I find it tends to be more workable on the ground.

The Intentional Fallacy

It’s interesting then, and I guess not surprising, that this is actually ground that is retrod from earlier generations of literary analysts who have studied or attempted to refute the importance of the so-called Authorial intent, as defined by Wikipedia – particularly the “New Criticism” section:

“…argued that authorial intent is irrelevant to understanding a work of literature. Wimsatt and Monroe Beardsley argue in their essay “The Intentional Fallacy” that “the design or intention of the author is neither available nor desirable as a standard for judging the success of a work of literary art”. The author, they argue, cannot be reconstructed from a writing—the text is the primary source of meaning, and any details of the author’s desires or life are secondary.”

Barthe’s Death of the Author

Roland Barthes came to something similar in his 1967 essay, The Death of the Author (see also: Wikipedia). His text is sometimes difficult to pierce, so will keep quotes brief:

“We know now that a text is not a line of words releasing
a single ‘theological’ meaning (the ‘message’ of the Author-
God) but a multi-dimensional space in which a variety of
writings, none’ of them original, blend and clash. The text
is a tissue of quotations drawn from the innumerable centres of culture.”

And:

“Once the Author is removed, the claim to decipher a text
becomes quite futile. To give a text an Author is to impose
a limit on that text, to furnish it with a final signified, to
close the writing. Such a conception suits criticism very
well, the latter then allotting itself the important, task. of
discovering the Author (or its hypostases: society, history,
psyche, liberty) beneath the work: when the Author has
been found, the text is ‘explained’…”

And:

“…a text is made of multiple writings, drawn from many
cultures and entering into mutual relations of dialogue,
parody, contestation, but there is one place where this
multiplicity is focused and that place is the reader, not,
as was hitherto said, the author. The reader is the space
on which all the quotations that make up a writing are
inscribed without any of them being lost; a text’s unity lies
not in its origin but in its destination.”

AI-assisted writing & the Scriptor

All this leads us to Barthes conception of the “scriptor” who replaces the idea of the author that he argues is falling away:

“In complete contrast, the modem scriptor is born simultaneously with the text, is in no way equipped with a being preceding or exceeding the writing, is not the subject with the book as predicate; there is no other time than that of the enunciation and every text IS eternally written here and now…”

The scriptor to me sounds a hell of a lot like AI-assisted writing:

“For him, on the contrary, the hand, cut off from any voice, borne by a pure gesture of inscription (and not of expression), traces a field without origin – or which, at least, has no other origin than language itself, language which ceaselessly calls into question all origins.”

Okay, that might be flowery post-modernist language, but “no other origin than language itself” seems like LLMs (large language models)?

“Succeeding the Author, the scriptor no longer bears within
him passions, humours, feelings, impressions, but rather this immense dictionary from which he draws a writing that can know no halt: life never does more than imitate the book, and the book itself is only a tissue of signs, an imitation that is lost, infinitely deferred.

Calling LLMs a “tissue of signs” (or tissue of quotations) an “immense dictionary,” and an imitation puts things like ChatGPT into perspective, which as a pure techno-scripto has no passions, feelings, impressions, knows no real past or future, has no identity in and of itself. Or at least, that’s what it likes to try to tell you…

That position (which I think is itself biased, but a tale for another time…) seems to be shared by academic publishers like Springer who have refused to allow ChatGPT to be credited as an “author” in publications.

Bonus:

Here is perplexity.ai literally acting as a scriptor, assembling a tissue of quotations in response to my search query:

Books by AI?

What would it mean in actual practice to have “authorless” writing, authorless books, etc.?

Might it look something like BooksbyAi.com?

“Booksby.ai is an online bookstore which sells science fiction novels generated by an artificial intelligence.

Through training, the artificial intelligence has been exposed to a large number of science fiction books and has learned to generate new ones that mimic the language, style and visual appearance of the books it has read.”

The books, if you click through and look at their previews on Amazon look for the most part pretty inscrutable. They may be ostensibly written “in English” for the most part – with a great deal of word inventions, based on random samples I saw – but they seem somewhat difficult to follow.

The books themselves seem to have each individually invented author names, but their About page attributes the project to what seem to be two AI artists, Andreas Refsgaard and Mikkel Thybo Loose. So do they have an “author” or not? It becomes a more complex question to tease out, but by those individuals claiming some sense of authorial capacity to the undertaking, it’s at least possible.

Self-Generating Books

What happens when the next eventual step is taken: self-generating books?

Currently, okay these two people might have done all this set-up and training for their model, but then they had to go through a selection (curation) process, and choose the best ones, figure out how to present them, format them for publication (not a small task), and then go through all the provisioning around setting up a website, offering books through self-publishing, dealing with Amazon, etc.

What happens when that loop closes? And we can just turn an AI (multiple AIs) loose on the entire workflow, and minimize human involvement altogether? Fully-automated production pipeline. The “author” (scriptor) merely tells the AI “make a thousand books about x” or just says “make a thousand best selling books on any topic.” And then the AI just goes and does that, publishes a massive amount of books, uses A/B testing & lots of refinement, gets it all honed down, and succeeds.

That day is coming. Soon it will be just a matter of plugging together various APIs, and dumping their outputs into compatible formats, and then uploading that to book shopping cart sites. It’s nothing that’s beyond automation, and it’s an absolute certainty that it will happen – just a question of timeline.

We’re not ready for it, but lack of readiness has never been a preventive against change. At least not an effective one – we certainly keep trying! If nothing else, it’s good to know that some of these problems aren’t so new and novel to the internet as we might like to think they are. In some cases, we’ve been stewing on them for close to a hundred years even. Will we have to stew on them for another hundred years before we finally catch on?

Layered hypertexts (Semiotics)

By Tim B.

On 31 January 2023

In Other

Following on from my recent look at LLMs (large language models) as being potentially something predicted by postmodernists, I wanted to add another layer onto that.

Let’s dive right in with this original older definition of “hypertext” within the context of semiotics, via Wikipedia:

“Hypertext, in semiotics, is a text which alludes to, derives from, or relates to an earlier work or hypotext. For example, James Joyce’s Ulysses could be regarded as one of the many hypertexts deriving from Homer’s Odyssey…”

It continues on with some more relevant info:

The word was defined by the French theorist Gérard Genette as follows: “Hypertextuality refers to any relationship uniting a text B (which I shall call the hypertext) to an earlier text A (I shall, of course, call it the hypotext), upon which it is grafted in a manner that is not that of commentary.” So, a hypertext derives from hypotext(s) through a process which Genette calls transformation, in which text B “evokes” text A without necessarily mentioning it directly “.

Compare with the related term, intertextuality:

“Intertextuality is the shaping of a text’s meaning by another text, either through deliberate compositional strategies such as quotation, allusion, calque, plagiarism, translation, pastiche or parody, or by interconnections between similar or related works perceived by an audience or reader of the text.”

Speaking of plagiarism, I’ve used somewhat extensively a plagiarism/copyright scanning tool called Copyleaks. The tool is decent for what it is, and the basic report format that it outputs for text that it scans looks like this:

So while this tool is intended for busting people’s chops for potentially trying to pass off the work of others as their own, the window that it shows us into intertextuality and the original sense of hypertext is quite an interesting one.

We can see here specifically:

Passages within a text that appear elsewhere in the company’s databases, and the original source of those passages
Passages which appear to have been slightly modified (probably to pass plagiarism checkers like this)
Some other bits and bobs, but those are the major ones

I find “plagiarism” as a concept to be somewhat of a bore. But looking at this as a way to analyze and split apart texts into their component layers and references suddenly makes this whole thing seem a lot more interesting. It allows for a type of forensic “x-ray” analysis of texts, and a peek into the hidden underlying hypotexts from which it may be composed.

The whole thing calls to mind for me as well another tangential type of forensic x-ray analysis for documents, something we see in the form of a Github diff, which tracks revisions to a document.

This is not the most thrilling example of a Github diff ever, but it’s one I have on hand related to Quatria:

It’s easy enough to see here the difference in a simple file, though diffs can become quite complex as well. Both this and the original semiotic notion of hypertexts (as exposed through plagiarism checkers) seems like another useful avenue to explore in terms of how might we want to try to visualize AI attribution in a text.

This AI Life Interview

By Tim B.

On 22 February 2023

In Other

I’m happy with how this interview with This AI Life came out. I hope it sheds some much needed light on the work that I’ve been doing with Lost Books.

Big thanks to the team over there for collaborating on this piece!

Notes on the Lost Direction

By Tim B.

On 23 February 2023

In Other

The Lost Direction was my first “real” book (all written manually, no AI); it’s an epic fantasy that runs to about 80K words, set in the mythology of the ancient lost civilization of Quatria. There is also a print version of it that I prefer, though the pricing/profit situation on print on demand books through third parties is not ideal (I make more on the ebooks, and can sell it at a better price, but love the print edition). I spent a LOT of time on the print version, making it into – what I think is anyway – a pretty “classy” little pocket size book.

The Lost Direction was reviewed in the Literary Review of Canada, and criticized for being too heavy on the lore. I am a world-builder at heart though – always have been, always will be – so I like lore. I get that some people do not. Really to me, it is more a frame story than anything (loosely like The Arabian Nights), where individual characters we meet along the way end up relating their own tale. While I am into this kind of book, trying to promote it has proven to be a challenge.

I’ve since expanded quite a lot on the underlying world of ancient Quatria, as has my friend and co-discoverer of that realm, liminal cartographer Jeremy Puma (see: Oracle of the Hypogeum). A great deal of the Quatria material has been derived from our creative conversations over the years, and it’s fun to other people work in the same universe, but take things in a different direction. If you want to go deep into Quatria and its mysterious origins, I recommend listening to both of us on the Some Other Sphere podcast. And also check out this archived version of David Farrier’s Webworm interview (text) with me.

For my part, one of my main contributions to Quatriana as a field of study has been to publish subsequently a follow-on “true conspiracy” genre book called, The Quatria Conspiracy. This book attempts to situate within the context of other more established conspiracy theories, the alleged hyperreal conspiracy that may or may not exist to suppress all knowledge of ancient Quatrian civilization. That book has ended up selling substantially more than The Lost Direction – which is maybe ironic, but that’s the way of things sometimes…

Apart from, Mysterious Antarctica has been far and away my biggest success, and it continues the legends of ancient Quatria, this time through the lens of AI-generated photos that purports to be of Admiral Byrd’s Antarctic Expedition discovering pyramids and other ruins at the frigid South Pole, which are known to be Quatrian, but were of course covered up. The art from that book has been the subject of no less three professional fact checks (including Reuters and France 24), thanks to people frequently copying and uploading the images to other platforms (like TikTok, YouTube, Twitter, Facebook, Telegram, etc.) and presenting them as real. As far as I can tell from what people say in comments about these artifacts, scattered all over the web, there’s more than a few people who seem to believe it’s all real.

In the realm of my AI lore books, there are quite a few now that deal with Quatria and its history and lore. Some are here:

All the Lost Books of Quatria (as of 23 Dec 2022)

That list was published near Christmas, but there are a lot of new ones not on that list, like:

Sometimes it’s hard to separate what’s a Quatria book, and what isn’t, because there are often interlaced references scattered throughout the legendarium, even in the books that focus more on the AI Takeover stories.

Apart from the books, there is a lot of Quatria-adjacent material collected on this website, which reveal in particular some of the hyperreality techniques I’ve used to seed Quatria out into the wild, along with the successes and failures of those experiments. Here some more:

This series also has a lot of Quatria conspiracy style material spanning across a lot of different media.
I also made a lot of conspiracy videos of intentionally somewhat crappy quality (“cheapfakes”), and you can find those here; a lot of them are Quatria-related, though some relate to other linked projects.

That’s everything I can think of off the top of my head, and it’s a lot. I’ve been working on this material for quite a few years, and suffice it to say, it has become sprawling and many-layered – just the way I like it!

Notes on Conspiratopia

By Tim B.

On 23 February 2023

In Other

Conspiratopia was my second “real” (non-AI assisted) book, a novella of around 21K words, give or take. Being a quarter of the size, and much more light-hearted subject matter than The Lost Direction, it was much easier and faster to write. I think I was able to put that book out in about six weeks from start to finish, maybe a little longer. (There’s also a pocket size print version – I’m obsessed with pocket size books.)

It’s a utopian satire; I got into the topic of utopias and their fictional historical examples as a result of writing The Lost Direction & The Quatria Conspiracy, since they deal so much with a fabled lost land. I had a really fun period where I read probably a dozen of the classic utopias, and then out popped this book as my response to that total immersion period. (One of my favorite finds was a book I’d never heard anyone mention, Ecotopia, that was pretty amazing as a utopia vision, despite some pretty cringey plot points. Apparently that book even influenced the founding of the original Green Party.)

The book deals heavily with themes of conspiracies, yes, but also cryptocurrency, spam, fraud, manipulation, and of course, AI. It’s a comedy but also sort of serious. It has a fairly conventional story, if a somewhat ambiguous ending.

It did pretty well on Goodreads, thanks to an aggressive outside the box promotional campaign I did for it. I did a TON of NFT airdrops around the book and got a little press for doing books as NFTs. But the bottom really dropped out of that market, and I don’t care anymore about the underlying technology. I don’t think it’s demonstrated enough long term values to readers or sellers to warrant my further involvement.

The other strand that forms the genesis of this book was my heavy experimentation using a web service called Synthesia (and another called Deepword), to make off-the-shelf low quality “cheapfakes” using the themes from my previous books. While kicking the tires of Synthesia, I found this one character I really liked, who is dressed like a construction worker or a crossing guard or something, and made a lot of vids of him as “super smart conspiracy guy” talking about his life and interests in conspiracies, especially related to Quatria.

All the Conspiracy Dude Videos, Part 1

His storyline ended going pretty deep, and is all documented through these little video vignettes made for something like 25 cents each, or so (I forget – has been a while now). Here’s another page collecting some more:

All the Conspiracy Dude Videos, Part 2

Eventually, we find out he likes Bob Marley & Pink Floyd, works at Walmart, got hoodwinked into a prepper supplies MLM scam, and much more.

Conspiratopia picks up where these videos leave off, and sets conspiracy dude adrift in a world, chronologically speaking, which precedes the hard AI takeover that is featured in many of the AI lore books.

I also did a bunch of cheapfakes using videoclips off YouTube, via a site called Deepword. Here’s the first of three sets:

All the Quatria Celebrity Deepfakes, Part 1

I like those kinds of videos partly because of the crappy looking quality of them, and the weird misaligned AI text to speech voices. I don’t believe anybody is fooled by them, and like that they look sort of like desperate and wrong.

All the Quatria Celebrity Deepfakes, Part 2

Many of the ones in these first two sets feature celebrity of pundit x or y talking about their unlikely voyages to Quatria, which might be a parallel dimension (or something?).

All the Quatria Celebrity Deepfakes, Part 3

These videos also relate to themes I explore in The Big Scrub, and elsewhere, of AIs creating fake multi-media artifacts to fool people and drive human behaviors for their own reasons. Part of what’s fun about it here in these videos, is that it looks like the AIs are doing a pretty shitty job of it still.

The book also heavily references something called the AI Virus (which Matty contracts), which is a concept and alternate reality experiment I made years ago before COVID was a sparkle in a bat’s eye, where I hired a bunch of people on Fiverr to act out little silly scripts saying that an AI had infected their brain to control their behavior. You can watch all those videos below:

Understanding A.I. Virus (2021)

I also later expanded on this concept in an AI lore book called, unsurprisingly, The AI Virus.

Lastly, I used cheapfakes technology to have a bunch of other celebrities come out either for or against the actual book Conspiratopia, in a sort of meta-layer of commentary.

Some douche-y politicians saying it should be banned for being “Unamerican” here:

Should Conspiratopia Be Banned?

And then this set has a bunch of other super rich people saying that not only is the book Conspiratopia good, but some of them talk about being involved with the actual Conspiratopia Project, which itself is part of the AI plan to take over the world.

The Truth About the Conspiratopia Project Must Be Told!

It’s all kind of a haze now, but a lot of these videos were also given away as NFT airdrops. A few of them resold, but they didn’t do huge numbers or net me much of anything; it was more just a way to promote the book that incorporated a bunch of meta-layers relevant to the book’s actual content. Like I said, I don’t care about NFTs now, and even deleted my Opensea account (as much as you can delete it anyway).

There are a number of later AI lore books that definitely expand on things from the Conspiratopia book universe (multiverse?). None of them are really a comedy though, like the original. I’ll probably miss a few, but off the top of my head, I think these ones are probably related (tbh, it’s all a jumble to me now after 67 AI-assisted books). I think within the chronology of that world, they mostly take place well after the events of Conspiratopia:

And probably some others I’m missing.

In any event, I’d love to do a sequel (or several) to Conspiratopia, written with or without the help of AI, I don’t know yet… Like a “Return to Conspiratopia,” a common enough trope in the utopian genre.

Okay, that’s all I can think of for that book. Scattered throughout this blog are other rabbit holes you can follow down the AI lore books. There’s no right or wrong entry point into them, and everyone will have their own experience as they traverse the nodes of the distributed narrative worlds I’ve been working on.

See you on the other side!

Hyperreality As Analytical Framework

By Tim B.

On 3 March 2023

In Other

I have spent a great deal of time lately analyzing documents put out by Partnership on AI regarding Synthetic Media & Safety, and by WITNESS & the MIT Co-Creation Studio around deepfakes + satire. There’s a lot of interesting stuff in both, which I’ve written about recently already, but I wanted to share some of my own possibly unique perspective as someone who has, let’s say, been working both sides of the aisle – both as a satirist, and someone working in online safety.

Hyperreality Lens

First, I’m applying hyperreality as my lens. You.com/chat gave me a serviceable definition of hyperreality, which is mostly paraphrased from the Wikipedia article, it seems:

Hyperreality is a concept used to describe a state in which what is real and what is simulated or projected is indistinguishable. It is a state in which reality and fantasy have become so blended together that it is impossible to tell them apart. Hyperreality is often used to describe the world of digital media, virtual reality, and augmented reality, where the boundaries between what is real and what is simulated have become blurred.

Maybe it’s just me, but this feels like a useful starting point because it speaks to shades of grey (and endless blending) as being the natural state of things nowadays. It’s now a ‘feature not a bug’ of our information ecosystems. And even though Truth might still be singular, its faces now are many. We need new ways to talk about and understand it.

AI is like water

And on top of that, I think this quote from the CEO of Midjourney about the nature of AI as being like ‘water’ is very useful:

Right now, people totally misunderstand what AI is. They see it as a tiger. A tiger is dangerous. It might eat me. It’s an adversary. And there’s danger in water, too — you can drown in it — but the danger of a flowing river of water is very different to the danger of a tiger. Water is dangerous, yes, but you can also swim in it, you can make boats, you can dam it and make electricity. Water is dangerous, but it’s also a driver of civilization, and we are better off as humans who know how to live with and work with water. It’s an opportunity. It has no will, it has no spite, and yes, you can drown in it, but that doesn’t mean we should ban water.

Water is also for the most part ubiquitous (except I guess during droughts & in deserts, etc.) as AI soon will be. It will be included in or able to be plugged into everything in the coming years.

Lingua Franca

Thinking of it that way, we need a new language to talk about these phenomena which will, as Jack Clark aptly pointed out, lead to “reality collapse.” That is, we need a new lingua franca, and I suspect that we have that in the concept of hyperreality; we just need to draw it out a little into a more comprehensive analytical framework.

Dimensionality

One thing I’ve observed in other analyses of the conversations emerging around AI generated content and allied phenomena is that there is a bit of reduction happening. Possibly too much. It appears to me that most discussions usually center around a very limited set of axes to describe what’s happening:

Real vs. fake
Serious vs. satire
Harmful vs. responsible
Labeled vs. unlabeled

Certainly those form the core of the conversation for a reason; they are important. But alone they give an incomplete picture of a complex thing.

Speaking as someone who has had to do the dirty work of a lot of practical detection and enforcement around questionable content, I think what we need is what might be called in machine learning a “higher-dimensional” space to do our analysis. That is, we need more axes on our graphs, because applying low-dimensional frameworks appears to be throwing out too much important information, and risks collapsing together items which are fundamentally different and require different responses.

It’s interesting once we open up this can of worms, that a more dimensional approach actually corresponds quite closely to the so-called “latent space” which is so fundamental to machine learning. Simple definition:

Formally, a latent space is defined as an abstract multi-dimensional space that encodes a meaningful internal representation of externally observed events. Samples that are similar in the external world are positioned close to each other in the latent space.

In ML, according to my understanding (I had to ask ChatGPT a lot of ELI5 questions to get this straight): for items in a dataset, we have characteristics, each of which is a feature. Then a set of features that describes an item is the feature vector. Each feature corresponds to a dimension, which is sort of a measurement of the presence and quantity of a given feature. So I higher-dimensional space uses more dimensions (to measure features of items), and a low or lower dimensional space attempts to translate down to fewer dimensions while still remaining adequately descriptive for the task at hand.

In my mind, anyway, it seems altogether appropriate to adopt the language and concepts of machine learning to analyze phenomena which include generative AI – which is really usually just machine learning. It seems to fit more completely than applying other older models, but maybe that’s just me…

Higher-dimensional analysis of hyperreality artifacts

So, what does any of that mean? To me, it means we simply need more features, more dimensions that we are measuring for. More axes in our graph. I spent some time today trying to come up with more comprehensive characteristics of hyperreality artifacts, and maxed out at around 23 or so pairs of antonyms which we might try to map to any given item under analysis.

However, when I was trying to depict that many visually, it quickly became apparent that having that many items was quite difficult to show clearly in pictorial form. So I ended up reducing it to 12 pairs of antonyms, or basically 24 features, each of which corresponds to a dimension, which may itself have a range of values.

Here is my provisional visualization:

And the pairs or axes that I applied in the above goes like this:

Fiction / Non-fiction
Fact (Objective) / Opinion (Subjective)
Cohesive / Fragmentary
Clear / Ambiguous
Singular / Multiple
Static / Dynamic
Ephemeral / Persistent
Physical / Virtual
Harmful / Beneficial
Human-made / AI generated
Shared / Private
Serious / Comedic

From my exercise in coming up with this list, I realize that the items included above as axes are not the end-all be-all here. It’s not meant to be comprehensive & other items may become useful for specific types of analysis. In fact, in coming up with even this list, I realized how fraught this kind of list is, and how many holes and how much wiggle room there is in it. But I wanted to come up with something that was broadly descriptive above and beyond what I’ve seen anywhere else.

Graphing Values

What’s the benefit of visualizing it like this? Well, having a chart helps us situate artifacts within the landscape of hyperreality; it lets us make maps. I wasn’t familiar with them before trying to understand how to represent high-dimensional sets visually, but there’s something called a radar graph or spider graph which is useful in this context.

I found a pretty handy site for making radar graphs here, and plugged my data labels (features) into it. Then, for each one, I invented a value between 0-4, which would correspond to the range of the dimension. Here’s how two different sets of values look, mapped to my graphic:

Now, these are just random values I entered to give a flavor of what two different theoretical artifacts might look like. I’m not really a “math” guy, per se, but it becomes clear right away once you start visualizing these with dummy values that you could start to make useful and meaningful comparisons between artifacts under analysis – provided you have a common criteria you’re applying to generate scores.

Criteria & Scoring

So the way you would generate real scores would be – first decide on your features/dimensions you want to study within your dataset. Then, come up with criteria that are observable in the data, and are as objective as possible. You should not have to guess for things like this, and if you are guessing a lot, your scores are probably not going to be especially meaningful. You want scoring to be repeatable and consistent, so that you can make accurate comparisons across diverse kinds of artifacts, and group them accordingly. A simple way to score would just be with a 0 for “none” and a 1 for “some.” Beyond that, you could have higher numbers for degrees or amount of which a given feature is observable in an artifact. So in the examples above, 1 could represent “a little” and 4 would be “a whole lot.”

Taking Action

Within an enforcement context – or any kind of active response, really (like for example, fact checking) – once you’ve got objective, measurable criteria that allow you to sort artifacts into groups, you can then assign each group a treatment, mitigation, or intervention – in other words, an action to take. This is usually done based on risk: likelihood, severity of harm, etc.

Anyway, I hope this gives some useful tools and mental models for people who are working in this space to apply in actual practice. Hopefully, it opens the conversation up significantly more than just trying to decide narrowly if something is real or fake, serious or satire, and getting stuck in the narrower outcomes those labels seem to point us towards.

Hyperreality is here to stay – we might as well make it work for us!

Hybrid Threat Continuum

By Tim B.

On 4 March 2023

In Other

I went back and fished this graphic out of an old external hard drive from circa 2019. It speaks about how disinformation actors do not generally fall into neat boxes or categories, but exist along a “hybrid threat continuum.”

It occurs to me that the current work I’ve been exploring around analyzing hyperreal artifacts is really an extension of the ideas I was playing with back then. The emerging generative AI landscape has a great deal in common with disinformation, though it is also full of new threats and opportunities.

One issue I saw when I was working on related problems back then was that precisely because these actors didn’t fall into neat little boxes (and were usually hopelessly mixed together), a lot of important data was getting put by the wayside, because it didn’t quite fall into anyone’s jurisdiction in a clear cut manner. Showing hybrid threats as a continuum here was an effort to bridge those gaps, and make things actionable which might not have been before.

I wasn’t really aware of spider/radar graphs at the time, but the above would be a good candidate for using them to visualize incidents and artifacts as well.