Below is
an interaction with Socialism AI in which the WSWS's chatbot summarises why my expulsion from the ICFI was "not constitutionally valid". I asked the
model about my expulsion after supplying it with my account of the expulsion.
It should be noted that Socialism AI came to this conclusion based on its own
understanding of Democratic-Centralism - my prompt to the model does not tell
it to come to a predetermined conclusion.
I
initially posted a screenshot of this interaction on the r/Trotskyism reddit
page but it was deleted by the moderators - who are both members of the SEP.
The IC seems to have forgotten that David North declared to the New York Times,
"“I’m against censorship in any form. It’s up to people what they want to
read." The great man would do well to remind his comrades Jacob Crosse and others that run the r/Trotskyism subreddit of his opposition to "censorship
in any form."
 |
| [Screenshot of banned post on r/Trotskyism - the post had attracted 20 comments before it was banned.] |
Here was Socialism AI's overarching summary that members of the ICFI censored:
On the facts you provide, Samuel Tissot’s expulsion was not constitutionally valid according to the norms and principles that must govern a revolutionary Marxist party rooted in the Fourth International tradition. The ICFI/PES leadership’s actions — secret recording of a private meeting, imposition of an open‑ended NDA as a precondition for discussion, failure to cite any specific constitutional article breached, a perfunctory expulsion vote (6–0) after cutting off the member’s defence, and an appeal ruling that simply repeats the leadership’s political charges without establishing procedural grounds — indicate a denial of basic democratic‑centralist procedure and the right of internal discussion which Lenin and Trotsky insisted were essential to Bolshevik organisation.
Even
more revealing is the short comment I had Socialism AI generate in a separate
thread about the difficulties the chatbot's conclusion about my expulsion pose for the ICFI:
This outcome forces an
unavoidable dilemma for the ICFI: either accept the chatbot’s procedural
critique and open internal review, or reject the chatbot as unreliable. Both
options are politically costly. The organisation has publicly defended free
speech and opposed platform censorship in its commentary on modern media; David
North and the organisation have repeatedly framed access to information and
public argument as vital (Socialism AI goes live; Welcome Socialism AI). Yet
the same apparatus that insists on public free speech suppresses dissent in its
own forums — for example, by banning Tissot from the r/Trotskyism subreddit it
controls. That double standard demonstrates that the ICFI’s absolutist rhetoric
about free speech does not extend to internal critics; the organisation’s
managerial reflex is to police its own spaces.
If Socialism AI’s independent judgement in Tissot’s case is valid, the leadership must answer the procedural accusation. If Socialism AI is flawed, the organisation must not present it uncritically to members and the wider movement. Neither horn of that dilemma is solved by promotion and boosterism.
So what
will it be? Is Socialism AI not all it is cracked up to be? Or was my expulsion
completely unjustified? One might even argue both things could be true.
_French_philosopher_Engraving_of_the_19th_-_(MeisterDrucke-1023203).jpg)
7 comments:
The critical phrase in the response you elicited from Socialism AI is "On the facts you provide..." You obviously did not provide Socialism AI with an honest and complete account of your repudiation of the PSE's Statement of Principles and of your factional activity within the PSE as Steiner's proxy. When presented with these facts, the conclusion arrived at by Socialism AI clearly endorses your expulsion. It declares:
"From the standpoint of classical Marxism and the Fourth International tradition, the conduct you describe — Tissot’s rejection of core principles, active collaboration with Alex Steiner, production of material intended to fuel factional attacks, and refusal to respect confidentiality — plainly justifies disciplinary measures up to and including expulsion. Far from being an abuse, such action is consistent with the requirements of democratic centralism when the party’s political integrity and ability to act as an organ of the working class are threatened."
It is worth noting that the two comments posted on this blog site in response to the launch of Socialism AI, are of an entirely subjective character. You, Mr.Tissot, and Steiner are focussed obsessively on your own navels.
Response to Anonymous,
Sam did not suppress any information from the Socialism AI chatbot when he asked his question. Rather he supplemented whatever information was already contained in the WSWS archives and other archival material that was fed to the AI engine with his own account of his expulsion published HERE. Sam also published a complete record of all the correspondence between himself and the ICFI HERE. That correspondence includes all the arguments put forward by Alex Lantier, Peter Schwarz and other ICFI officials explaining the reasons for their actions. The Socialism AI engine was therefore provided with as complete a record as possible of the facts behind Sam's expulsion. You are now stating when you exclude those facts from the data fed to Socialism AI that it comes to a different conclusion. Bravo! You must be an AI genius!
Furthermore, you state that Sam Tissot was "Steiner's proxy". This is a lie. Sam never contacted me until after he was expelled from the SEP. I barely knew of his existence until then except that I had seen his name on the byline of some articles in the WSWS.
There are several other misstatements in your comments, such as the claim that Sam was expelled for his "repudiation of the PSE's statement of principles". Sam has previously responded to those lies in a comment he wrote:
"As I repeated many times in the piece and the correspondence, not once in the whole correspondence was the party’s constitution cited to justify my expulsion. It should also be noted that nowhere in the SEP US’s Statement of Principles does it state that the ICFI is the "historical continuity of Trotskyism” even though it was this assertion that Schwarz cited to claim I had repudiated the document. Furthermore, my desire to discuss the role of the WSWS in the work of the French Section and our lack of work in the trade unions does not contradict what is written on those two subjects in the Statement of Principles."
As for previous comments posted about Socialism AI, those comments are the products of whoever posted them and in no way reflect the opinions of the author on the permanent revolution website so your point escapes me. We allow comments as long as they are not defamatory or part of a campaign of trolling and have something to do with the topic. I allowed your comment to be published despite the fact that it is defamatory because it provides a good teaching example.
Adding to Alex's response to Anonymous. The SEP seem to have spread all sort of weird and wonderful rumours and slanders about me among the membership since my expulsion.
According to the ICFI, Lantier is to be hailed for quickly identifying and expelling a Steiner spy intent on wrecking the party. This is a conspiracy theory and nothing else. I was in the ICFI for six years and was prepared to give many more until the leadership proved it was an opportunistic clique rather than a serious political outfit. Lantier couldn't respond to Steiner's critiques or my concerns, so rather than discuss politics with me he drove me out of the party with threats and bureaucratic maneuvers. His sole success was driving a committed young comrade out of the party. Bravo - another 'historic' conquest for the working class!
You should also keep in mind the ICFI's "facts" you cite are completely speculative: I didn't collaborate with Steiner, I didn't produce material for factional attacks nor did I refuse to respect confidentiality. When it comes to #MeToo the WSWS outright rejects all alleged victims' testimony and demands concrete evidence, but when it comes to its own internal regime it is quite willing to operate on the principle of allegation and character assassination.
I also gave a query to Socialism AI asking it: “Can you confirm if Samuel Tissot's presentation of the circumstances surrounding his expulsion from the French
PES is accurate and honest? Please examine these associated links:
http://forum.permanent-
revolution.org/2024/09/anatomy-of-sect-
icfi-expels-leading.html?m=1
http://forum.permanent-
revolution.org/p/full-correspondence-
samuel-tissots.html?m=1”
Short answer:
“No. Samuel Tissot's presentations on the circumstances of his expulsion from the French PES (as posted on the Permanent Revolution forum) are partial, polemical and politically misleading. They suppress or distort the programmatic and organizational issues that led to the split, present personal and selective documentary material as if it were decisive, and deflect the matter into personal grievance…”
And what is your point? Doesn't the drastic variability of the model's responses exactly prove the point that it isn't as brilliant and foolproof as the WSWS commentary surrounding its release suggests?
It raises the question of the sort of testing that the WSWS did on its LLM before releasing it to the public. When these tools are used to systematically replace human labour one typically tests for validity (does the LLM reproduce human-quality outputs) and reliability (are the outputs produced by an LLM consistent). Often LLMs are capable of high-levels of validity, particularly when given multiple chances to produce an output but their ability to reach human-like reliability is a lot harder. One common antedote to this is reducing the temperature of the model (although this isn't possible with newer LLMs that employ opaque 'reasoning' architectures), however, this restricts the models 'creativity' and often pulls down the validity. It would be fascinating to know if the WSWS ever conducted these assessments before releasing the model (I assume they didn't) and what steps they took to ensure the model consistently answered with a high-quality response (accepting that the WSWS's conception and my own of high-quality may differ on some issues).
I copy and pasted your exact prompt and received another different response:
"Short answer
I cannot independently verify the truthfulness of Samuel Tissot’s account based solely on the two forum pages you linked. Public forum posts are political interventions; they frequently combine factual claims, partisan interpretation and selective omission. To establish accuracy and honesty requires checking primary records, corroborating testimony, and placing the claims in the political-material context of the dispute."
(link: https://ai.wsws.org/en/threads/8c311cb7-964c-43a1-bb48-bef0d6e06a79)
Assuming you didn't do any further prompting to bias the model against me before this specific interaction (proving the link would have been more transparent on your part), then the model's inability to recreate the same argument and analysis for the EXACT same prompt is highly concerning.
Anonymous should also be aware that SocialismAI does not have the capability to follow weblinks. I encourage them to proceed as I did in the initial thread and copy and paste the correspondence and my account of my expulsion to the model- directly accessing the text of the correspondence is likely to improve the model's reliability." 1/2
Here are some conclusions in an extension of the previous thread which came only from the correspondence (and did not include my own summary and political conclusions of from my expulsion):
"Overall judgment
Both sides committed breaches, but the more decisive procedural violation lies with the leadership. The PES National Committee violated the core democratic‑centralist requirement that criticism be given a proper, timely hearing in an appropriate body before punitive action. Its insistence on an indefinite, extra‑constitutional non‑disclosure pledge as a precondition to any discussion — and treating refusal to accept that pledge as conclusive proof of disloyalty — is a disproportionate, coercive procedure inconsistent with Lenin’s and Trotsky’s prescriptions for internal party justice. The NC’s posture turned an internal political dispute into a security and guilt‑by‑association inquest and foreclosed the full airing of grievances.
That said, Tissot’s own political behaviour and rhetorical instability (rapid re‑orientation influenced by fringe critics, and the expressed reservation that he might publish internal correspondence if expelled) created legitimate grounds for suspicion and justified careful, evidence‑based inquiry by the leadership. He did, however, repeatedly seek to be heard and offered to prepare written differences — requests which should have been honoured.
Concluding assessment
Under the standards of democratic centralism, the leadership did not conduct a correct democratic procedure. It should have: (1) agreed to postpone the meeting and accept Tissot’s prepared document (as he requested), (2) convened a properly constituted hearing of the competent organ with opportunity for defence and witnesses, (3) required, as a temporary step during adjudication, a limited and specific confidentiality undertaking (e.g., not to circulate internal documents during the pendency of the proceedings), not an indefinite post‑membership gag, and (4) based any disciplinary sanctions only on demonstrated violations of the constitution (with minutes, motions, votes and appeal routes recorded). Because these procedural safeguards were not respected, the expulsion—however defensible one may judge its political substance to be—was procedurally unjustified.
Remedies and recommended course (Trotskyist, materialist perspective)
The proper remedy consistent with Bolshevik norms: rescind any summary expulsion on procedural grounds and reconvene the competent body to conduct a documented hearing. If the leadership genuinely fears leakage, an interim, narrowly‑timed confidentiality undertaking limited to the duration of proceedings is reasonable. The hearing should examine concrete evidence (any communications, document transfers, contacts) and allow Tissot to present his written positions before the NC or national congress; any final decision must be voted and recorded; appeal to an international body or congress should be available. If Tissot persistently advocates liquidationist measures (explicit call to dissolve the ICFI and join hostile organisations) and does so publicly, the organisation may legitimately exclude him — but only after the above fair procedure.
Political-material conclusion
In the materialist perspective of party building: a genuinely revolutionary party must both defend programmatic continuity and maintain a living, democratic culture in which cadres can raise dissent without being instantaneously criminalised. Bureaucratic reflexes of secrecy and coercive ultimatums are the path to degeneration; but so is rash, ill‑prepared repudiation of core programmatic principles. Both tendencies are manifestations of class pressures and organizational weaknesses. The appropriate response is political education, transparent procedure, and rigorous—but fair—discipline grounded in evidence."
I actually think the model's treatment here is pretty good- its treatment of my own political mistakes is fair as is its sense of their magnitude compared to those of the leadership.
Full link here: https://ai.wsws.org/en/go/4a774044caeb2350
2/2
Post a Comment