Jianwei Xun: Fake philosopher
« previous post | next post »
Jianwei Xun, the supposed philosopher behind the hypnocracy theory, does not exist and is a product of artificial intelligence
A collaboration between an essayist and two AI platforms produced a book that reflects on new forms of manipulation
Raúl Limón, EL PAÍS (4/7/25)
The entire proposition behind this scheme is so preposterous and diabolical that I am rendered virtually speechless.
The French city of Cannes hosted a roundtable discussion on February 14 called “The Metamorphosis of Democracy – How Artificial Intelligence is Disrupting Digital Governance and Redefining Our Policy.”
The debate was covered in an article by EL PAÍS after Gianluca Misuraca, Vice President of Technology Diplomacy at Inspiring Futures, introduced the concept of “hypnocracy” — a new form of manipulation outlined in a book by Jianwei Xun called Hypnocracy: Trump, Musk, and the New Architecture of Reality. However, this Hong Kong philosopher does not exist, as revealed by Sabina Minardi, editor-in-chief of the Italian magazine L’Espresso.
There is a real book out there. It has sold thousands of copies. It has occasioned vigorous debate. It is having consequences. Yet, although the idea behind it was created by an Italian publisher and bookstore owner named Andrea Colamedici (b. 1987), it was authored by two AI platforms, i.e., non-humans.
The French city of Cannes hosted a roundtable discussion on February 14 called “The Metamorphosis of Democracy – How Artificial Intelligence is Disrupting Digital Governance and Redefining Our Policy.”
The debate was covered in an article by EL PAÍS after Gianluca Misuraca, Vice President of Technology Diplomacy at Inspiring Futures, introduced the concept of “hypnocracy” — a new form of manipulation outlined in a book by Jianwei Xun called Hypnocracy: Trump, Musk, and the New Architecture of Reality. However, this Hong Kong philosopher does not exist, as revealed by Sabina Minardi, editor-in-chief of the Italian magazine L’Espresso.
In this case, "Jianwei Xun", which is a surrogate for two non-human agents and one human idea man, is different from a pseudonym, because it is the composite representation of a semi-human, semi-machine entity. Who / what is responsible for the book and what it entails?
The theory was created by essayist and editor Andrea Colamedici, who, although listed as a translator, is actually the co-author of the book, along with two AI platforms. This fact was never disclosed, violating EU law on AI. In response, EL PAÍS decided to remove the article published on March 26, which referenced content from the book attributed to the non-existent Hong Kong philosopher.
How was the machination exposed?
Journalist Sabina Minardi, after multiple attempts to interview the supposed author, uncovered that it was a pseudonym for a work created and edited by Colamedici in collaboration with two AI tools, part of a project to reveal AI’s influence in producing coherent and convincing discourse. As a result, researcher [Cecelia] Danesi referred to this new form of manipulation as a “digital dictatorship.”
If you take a look around the internet for the keywords "hypnocracy", "Jianwei Xun", etc., you will see that they have proliferated and are having an impact. But who is morally and legally responsible for their consequences? Has authorship been falsely attributed? If so, is it subject to criminal prosecution? Have someone's moral rights been abrogated or trespassed? Has anyone been defamed? Has fraud been committed?
The EL PAÍS article closes with a defense of Hypnocracy as a worthwhile "experiment" in discussing a complex of thorny issues facing society and a more nuanced approach to the phenomenon than the one with which I began this post. The human author / actor pleads that he did not intend to deceive anyone, only wished to raise critical arguments. Yet, Jianwei Xun has a website, and his / its legacy lives on….
Selected reading
- "Trespassed update" (10/15/24) — with relevant bibliography
- "Hypnocracy" — YouTube video (8:42) in which the human author calls himself part of a "collective"
[h.t. Bryan Van Norden]
Chris Buckey said,
May 18, 2025 @ 1:22 am
Uggggghhhhhh
Peter Grubtal said,
May 18, 2025 @ 2:26 am
Nobody should be surprised. The stuff of much of contemporary sociology and philosophy is just pretentious content-less verbiage and it didn't take AI (Sokal affair)to demonstrate that it's easy to produce rubbish that will pass muster in those circles.
AntC said,
May 18, 2025 @ 4:28 am
Every AI-generated text longer than a few phrases I've come across 'smells' not written by a human. Did the translators not notice the source text seemed dubious?
If not, then isn't the text really human-authored by the 'essayist' (Andrea Colamedici)? That they used AI-generated material would be no different to (say) assembling a book by cribbing articles from wikipedia(?) There's many programmer tutorials churned out by this means. They're mostly worthless for learners; you get a 'nose' for the style.
Does the book include cites/references/name names of experts? (Something AI texts are notoriously weak at.) Do they point to actual authors and publications? Did those "intellectuals and philosophers" discussing the book not carry out usual academic due diligence?
Or is the whole brouhaha (including the Cannes 'roundtable') no more than journalistic spruiking the book?
David Marjanović said,
May 18, 2025 @ 6:14 am
That was thirty years ago, though.
Most likely.
Peter Cyrus said,
May 19, 2025 @ 3:37 am
Jianwei Xun … does not exist.
So Lewis Carroll and Mark Twain didn't exist, either. I'm not clear on what is outraging some of us: the use of a pseudonym, collaborative writing, the participation of AI, or the intent to deceive (or at least conceal).
I have a cousin who used to perform on stage, and chose a stage name, perhaps because it sounded more interesting, and perhaps also to draw a line between her professional and personal lives. But curiously, after many years, she changed her legal name to her stage name, and the old name is now forgotten. I don't see anything nefarious in that.
Peter Taylor said,
May 19, 2025 @ 4:49 am
The only reading of this which makes sense to me is that the objection to the word pseudonym is that the name conceals a composite authorship, but I don't really understand the objection. The phrase collective pseudonym is not a novel coinage and in practice the qualification is often elided. I'm reminded of the 2021 scandal over Carmen Mola, an apparently female author who, when they arrived to collect a literary prize, turned out to be three men. I think all of the Spanish sources I've seen on the subject call the name a pseudonym, and the Spanish Wikipedia page today certainly does. (The English one explicitly says collective pseudonym).
Victor Mair said,
May 19, 2025 @ 5:54 am
The participation of AI makes it different from a pseudonym.
bks said,
May 19, 2025 @ 5:55 am
Encyclopedia Britannica lists Asclepius as one of the "100 Most Influential Scientists of all Time: https://www.arvindguptatoys.com/arvindgupta/hundred-scientists.pdf
This despite the fact that Asclepius is not a mortal man: https://en.wikipedia.org/wiki/Asclepius
Victor Mair said,
May 19, 2025 @ 6:09 am
A god is different from AI, and vice versa.
Rodger C said,
May 19, 2025 @ 11:57 am
A god is different from AI, and vice versa.
There's a kind of Heraclitean ring to that.
Philip Anderson said,
May 19, 2025 @ 1:56 pm
@Peter Cyrus, Peter Taylor
Pseudonym: when does a false name become a fake identity?
Many writers have concealed their names for various reasons:to keep their their literary life (or lives) separate from their occupation or private life, or to protect their life, career or social status. For some literary competitions, such as Welsh eisteddfods, it is compulsory.
But a name may also mislead the reader, as when a woman uses a male pseudonym, or vice versa. That has a long history, e.g. the Brontë Sisters, often in order to be taken seriously. Does it matter? I don’t think so, unless the writer explicitly claims to be writing from personal experience, which should exclude fiction.
But choosing a name that falsely suggests a particular minority heritage is controversial, and generally unpopular with minorities, who feel exploited (again). I know there are people who disagree that minorities, or any vulnerable people, deserve particular protection. So a fake Hong Kong philosopher rings alarm bells.
A pseudonym may also conceal multiple authorship; again is that a problem? There are husband and-wife partnerships who write detective novels. For competitions of course, it may be against the rules, as happened in one National Eisteddfod.
But an author is still assumed to be a living being, generally human although books have been attributed to gods; with AI , that is not necessarily the case. In the EU, that deceit is apparently illegal. It certainly poses copyright questions, as would any kind of automatic writing.
For fiction, perhaps we can’t expect to learn anything about the author, and the text should be read for what it is. Plenty of authors have written a fake frame narrative around the main story, claiming to have discovered an ancient manuscript. For a work that really claims to be true, it may be a different matter.
Gregory Kusnick said,
May 19, 2025 @ 5:24 pm
The El Pais article says straightforwardly that "The theory was created by essayist and editor Andrea Colamedici" and quotes Colamedici's claim that the book "said things that I myself believe in and have produced." So it seems reasonable to conclude that for practical purposes, Colamedici is the author and Jianwei Xun is his pseudonym. The fact that he used AI tools (not "agents" or "collaborators") to produce the text seems irrelevant to the question of attribution.
But let's look ahead a few decades to a time when advanced AI systems are capable of producing book-length, human-level works of scholarship entirely on their own initiative, without human intervention. In such case, it seems to me, an AI author would have as much right to publish pseudonymously as anybody else. Let the work speak for itself, regardless of who (or what) wrote it.
Victor Mair said,
May 19, 2025 @ 8:11 pm
We are posing profound questions here.
What happens if the machine-author becomes a diabolical demagogue and successfully incites mobs to destructive violence? Who will be held accountable? What if the machine-author gains control of a robot army and drone air force? Did you see the Chinese viral video of a robot that went out of control? That was a minor event, but things could get much worse in the future when the AI-mind machine interface becomes much more powerful?
Science fiction authors used to write about such scenarios, and we used to laugh them off because they were merely imaginary fiction. But AI cortices have already taken on some of the properties of an independently cogitating mind whose interactions with humans and machines have consequences in the real world.
I think the EU was right to put legal restrictions on what can be done with such a "thinking" mechanism.
Roscoe said,
May 20, 2025 @ 9:30 pm
Rodger and Prof. Mair:
“It takes one to know one, and vice versa!” – Alfred E. Neuman
Chris Buckey said,
May 21, 2025 @ 5:43 am
There's almost nothing profound about large language model-based AIs. They're not even artificial intelligence in any meaningful way, just glorified autocomplete functions that waste incredible amounts of electricity and hardware capacity.
Benjamin E. Orsatti said,
May 21, 2025 @ 8:11 am
Chris,
It gets me too how rarely the energy angle comes up. Like, maybe we should cool it on the whole AI thing until we make some appreciable progress on a Dyson Sphere?
Chris Buckey said,
May 21, 2025 @ 9:17 am
Benjamin,
Any civilization capable of building a Dyson Sphere (assuming such a structure is possible from a constructional standpoint) has probably come up with something far more clever and useful than LLMs already.