Bad Habits: Institutional Reality Erasure 🤷

Bad habits look good on me
Little bit wrecked
But I wear it clean
Lipstick stain on a white tee
Still walking like a silver screen
— Bad Habits Look Good On Me — VAELYN

Tuesday, August 25th 2026

I published a thing.

The title:
"Slutty Cybernetics: Two Cyberneticists. One Meeting. 🍵"
(Yes, you're reading that right. 😉)

Then I edited that thing into a boosted LinkedIn post. The boost got rejected.

I chose to post calmly about that:

(The boost for that one went through. Which is the thesis demonstrating itself in real time. On LinkedIn. On a Wednesday morning, August 26th, 2026. Just observing.)

I also chose to use LinkedIn's premium Support Chat. An AI responded. I requested calmly to speak to a human and pointed out that this was about consent.

AI: "Hi there. With the help of AI, I can answer questions about your Premium Account or connect you to our team. Not sure where to start? You can try: [irrelevant list]"

(No human in the loop yet.)

Alex: "Hi, I would like to challenge a non-approved ad on the grounds that the suppression is itself a diagnostic of consent violation. I'm a cyberneticist. I'm a communications researcher. I can quote Watzlawick if you want. Can I please talk to a human?"

AI: "Sure. I can transfer you to our team. Did you need support with one of the following products? [the same irrelevant list]"

Alex: "I wanna talk to a human. Thank you."

AI: "I can help you create a ticket and our team will answer your question as soon as possible."
AI: "Send your additional details in one message and we will share it with our team to follow up on."

Alex:
Hello dear LinkedIn team.

I'm requesting a formal statement on the non-approval of the following ad:
https://www.linkedin.com/posts/systemic-engineer_cybernetics-opentowork-tech-activity-7496451922921701376-E-7q

On which grounds was the ad not approved? And how does this constitute safety for your users? I presume the word "slutty" might have triggered a content filter. Considering that the formalization of the mathematics of consent are in large parts derived by the practices of the BDSM community (as laid out in the piece below) this consideration does not hold up under scrutiny.
https://systemic.engineering/p/4ffb4621-d54d-4be6-8b90-938c7ec2e8ee/

Care to elaborate?

With kind regards and waiting patiently for your answer
Alex 🌈

AI: Ticket submitted!

(No human in the loop yet.)

Wednesday, August 26th

LinkedIn replied!

Unnamed LinkedIn Support:
Hi Alex, I'm so sorry that I don't have a quick answer to your inquiry. Due to the sensitive nature of your request, I've forwarded this message to our Safety team. Please understand that your issue may require some additional research, which may delay their response. Someone from that team will get back to you as quickly as possible. If you're able to access your account, you can check the status of your case at any time from the Your cases tab on LinkedIn Help. Learn more here: https://www.linkedin.com/help/linkedin/answer/43378 Thanks for your patience as we look into this for you.

Warm Regards,
Premium Support Consultant

(Ruthvick enters the ticket.)

Ruthvick:
Hi Alex,

Thanks for reaching out. Your ad was not approved because it did not comply with LinkedIn’s Advertising Policies, which require all ads to be accurate, safe, and respectful. Our review flagged the ad for violating [specific policy area], such as:

Prohibited content (e.g., misleading claims, unsafe products, adult/violent content)
Restricted content without required approvals Inappropriate targeting or creative (e.g., discriminatory, sensational, or harassing language)
Brand safety concerns (e.g., links to untrusted destinations or insecure experiences)

These standards exist to protect members from misleading, harmful, or unsafe experiences and to maintain a trustworthy environment on LinkedIn. Applying these policies consistently helps ensure user safety, content integrity, and relevance across the platform. If you’d like, I can share the exact policy section(s) your ad triggered and guidance to revise it for compliance. Please send the ad ID or a copy of the creative and landing page, and I’ll outline the specific changes needed for re-submission.

Regards, Ruthvick LinkedIn Trust & Consumer Support Consultant.

Alex:
Hi Ruthvick,

Thank you for your timely answer. I expected silence. This is a pleasent surprise.

Please be so kind to elaborate on LinkedIn's concept of Safety, how the specific block enacted that Safety, who specifically that decision protected and who, by name, is responsible for the Safety design.

The research of Lore Born and I has shown very clearly that accountability goes to die through observer-exclusive language. That is citing policies without names and humans attached to them. I'm aware you have a playbook, Ruthvick. I'm calmly informing you that this is a case where the playbook does not apply.

If I come to conclusion that accountability is being avoided through playbook-observer-exclusive language then I'm hereby informing you that my Förster-canonical practice dictates I deconstruct the decision publicly on systemic.engineering. Including screenshots and verbatim quotes.

This is not a threat. This is a professional disclosure from a researcher that is specifically focussing on how institutional language erases the lived-experience of consent-capable subjects in automated and corporate decision making.

With kind regards
Alex 🌈

(Alex opens LinkedIn search. Searches for "Ruthvick" and working at "LinkedIn". Zero results.)

(No human in the loop yet?)

Ruthvick:
Hi Alex,

Thanks for reaching out to us about this. If you could, please provide me with as much of the following information as possible so I can investigate this:

• The name and/or email address of the member(s).
• A URL that links to the reported profile (e.g., the public profile URL found through an internet search or the URL resulting from a people search after you sign into your LinkedIn account).
• A link to the reported content. Learn how to obtain the link here: https://www.linkedin.com/help/linkedin/answer/108024
• A screenshot of the reported content. Learn how to capture and send us a screenshot here: https://www.linkedin.com/help/linkedin/answer/223

I look forward to hearing your response in order to further assist you.

Regards, Ruthvick LinkedIn Trust & Consumer Support Consultant.

Alex:
Hello Ruthvick,

Let me just quote the opening of my own ticket:

Hello dear LinkedIn team.

I'm requesting a formal statement on the non-approval of the following ad:
https://www.linkedin.com/posts/systemic-engineer_cybernetics-opentowork-tech-activity-7496451922921701376-E-7q

On which grounds was the ad not approved? And how does this constitute safety for your users? I presume the word "slutty" might have triggered a content filter. Considering that the formalization of the mathematics of consent are in large parts derived by the practices of the BDSM community (as laid out in the piece below) this consideration does not hold up under scrutiny.
https://systemic.engineering/p/4ffb4621-d54d-4be6-8b90-938c7ec2e8ee/

Care to elaborate?

With kind regards and waiting patiently for your answer
Alex 🌈

—
Anything else I can elaborate on, Ruthvick?

Alex:
Hello Ruthvick,

as stated before I felt obligated through my Förster-canonical practice to document this exchange with LinkedIn about consent and institutional reality erasure. This conversation is now a public metalogue.
https://systemicengineering.substack.com/p/bad-habits-institutional-reality

I recommend gently to consider how you'd like to respond to that internally. 🙏

With kind regards
Alex Wolf 🌈
(they/them)

Thursday, August 27th

Ruthvick:
Hi Alex,

I wanted to let you know that we've received your message, and we're currently researching this matter.

Thanks for your patience!

Regards,

Ruthvick
LinkedIn Trust & Consumer Support Consultant.

Ruthvick:
Hi Alex,

Thanks for your reply.
We noticed that the link included in your recent post appears to be broken. To ensure your audience can access the intended content, please edit the post and replace the existing link with the correct URL referenced in this ticket: https://systemic.engineering/p/4ffb4621-d54d-4be6-8b90-938c7ec2e8ee/.

Once the link is updated, the issue should be resolved. If you need any assistance with editing the post or have questions, feel free to reply to this email.

Alex:
Hi Ruthvick,

Thank you for the trouble shooting support.

Would LinkedIn be able to provide an answer to the questions I posed earlier:

What is LinkedIn’s definition of Safety?

How did the rejected ad boost specificly enact that Safety?

Who did that decision protect and how?

Who, by name and function, carries the ownership and accountability for the Safety design?

With kind regards

Alex Wolf

Alex:
And for completeness sake, here is a confirmed working link to the post whose ad boost got blocked:

https://www.linkedin.com/posts/systemic-engineer_cybernetics-opentowork-tech-activity-7496451922921701376-E-7q

Friday, August 28th

Ruthvick:
Hi Alex,

Thanks for your questions about Safety on LinkedIn. In brief:

What "Safety" means on LinkedIn: We work to keep conversations safe, trusted, and professional. Our Professional Community Policies set rules for what's allowed, and we enforce them using automated systems plus human review when needed.

How a rejected ad enacts Safety: All ads are reviewed against LinkedIn's Advertising Policies and editorial standards before they can run. If an ad or boost is rejected, it's because it's content, landing page, targeting, or format didn't meet those policies. You'll see the specific reason in Campaign Manager and can edit and resubmit or appeal from the notice page.

Who is accountable for Safety design: Safety is owned across dedicated LinkedIn teams — Trust & Safety (content integrity), Ads Abuse (ad policy enforcement), Ad Policy (policy design), Product-Ad Trust (automated approvals), and Legal - Digital Safety & Content (policy oversight). For privacy and security, we don't share individual employee names externally, but these functions collectively carry ownership and accountability.

Reference Links:

Professional Community Policies: https://www.linkedin.com/legal/professional-community-policies
How we enforce our policies: How we enforce our Professional Community Policies | LinkedIn Help
How LinkedIn keeps you safe: How LinkedIn Keeps You Safe | LinkedIn Help
Advertising Policies: https://www.linkedin.com/legal/ads-policyAd guidelines and review: Ad guidelines for ad review and approval | Marketing Solutions Help
Rejected ads help: https://www.linkedin.com/help/lms/answer/a416939

If you believe your ad already complies, please use the in-product appeal link shown on the rejection notice.
If you prefer to adjust and resubmit, the notice page includes a direct “Edit” link.

Regards,

Ruthvick
LinkedIn Trust & Consumer Support Consultant.

(Alex opens LinkedIn's Campaign Manager.)

Alex:
Hi Ruthvick! (Apologies for the misspelling earier. Dsylexic brain. 🤷)

Thank you for pointing me to Campaign Manager as the interface where the specific reason is shown. I investigated and would like to share what I found.

I opened the Campaign Manager and was greeted by the "Set up your first ad set"-wizard. The notifications were empty.

I documented this with screenshots on: https://systemicengineering.substack.com/p/bad-habits-institutional-reality

This either suggests that "ad boosts" created directly through the post are not managed by the Campaign Manager, or that there is a bug on your side. Which of those is it?

I've also reviewed the full LinkedIn Advertising Policies and Professional Community Policies. To my reading, none of the prohibited or restricted categories unambiguously apply to a boost of a fictional consulting case-study titled 'Slutty Cybernetics: Two Cyberneticists. One Meeting.' — a piece that names a real professional-methodology using a metaphorical brand-vocabulary.

Considering that I can not verify the block reason from my end through the resources you provided, let me restate the question: Which specific policy section was cited by the automated classifier at the moment of block? And what was the confidence-score / signal-type that fired? And who did that decision protect specifically?

Furthermore I'd like to understand what how your internal feedback loop to deal with false-positive classifications looks like specifically. You've listed five different teams that are responsible for the automated handling of ads on LinkedIn. How do your internal processes ensure that the ownership of any given feedback, like this one right now, is being properly metabolized by the team, and the people on the team, that actually have the authority to act on the feedback and alter the system under observation? (In this case LinkedIn's ad classifier.)

With appreciation for your continued collaboration, Ruthvick.
Alex Wolf 🌈

Der Koffer knurrt lauter. 🖤

Ruthvick:
Hi Alex,

Thanks for contacting Linkedin Support. I hope you are doing well. I'm sorry for not having a quick answer about your issue. I've forwarded your message to another group for additional review and advice. We'll be in contact with you as quickly as possible. Your issue may require additional research, which may extend your wait time. If you can log into your account, you can update and check the status of your case on the LinkedIn Help Center Your cases page: https://www.linkedin.com/help/linkedin/cases

Please note that if you can't log in to your account, you won't be able to check the status of your case. We ask that you don't create additional cases in the meantime. We're working as quickly as possible to resolve your inquiry. Thanks for your patience.

Regards, Ruthvick
LinkedIn Trust & Consumer Support Consultant.

Alex:
And in case it turns out that LinkedIn cannot answer these questions from within without admitting that there's no mechanism to enforce the feedback loop. Maybe you wanna consider hiring Two Cyberneticists for One Meeting. We're #OpenToWork and that's literally what we do for a living. 😉

(Irina entered the ticket.)

Irina:
Hi Alex,

Thank you for reaching out in relation to LinkedIn Marketing Solutions. Unfortunately, I don't have a quick answer to your request.

I've forwarded your inquiry to our LinkedIn Marketing team for assistance. Someone will be in contact with you soon regarding your inquiry. In the meantime, you may receive a second notification confirming that they've received your message.

In the meantime, you can search on the LinkedIn Marketing Solutions Help Centre:
https://www.linkedin.com/help/lms

You should be hearing from the LinkedIn Marketing Solutions soon. We look forward to working with you and appreciate your patience in the meantime.

Regards, Irina
LinkedIn Trust and Consumer Support Consultant

(Alex opens LinkedIn search, types “Irina”, selects “LinkedIn” in the Company filter, finds 3 results. That’s progress, I presume.)

Alex, from the fourth chair, reads Irina's response out loud:

We look forward to working with you [...]

Alex turns to Loki, who is polishing the bar with a cloth that's always dirty and never clean.

Alex: "How does that read to you, Loki?"
Loki: "Like the agent accidentally selected just the right template."

Alex nods. Takes a sip from the Barbaresco.

Alex: "Bemerkenswert."
Loki: "Bemerkenswert."

The Roomba bumps into the shelf with the LinkedIn logo, next to the pinboard. Adjusts. Moves forward. The shelf materialized a few days ago, which is what happens in Silicon Venue when the recursion stops moving.

Saturday, August 29th

LaRue entered the ticket.

LaRue:
Hi Alex,

Thank you for reaching out in relation to LinkedIn Pages. I've forwarded your inquiry to our LinkedIn Pages for assistance. Someone will be in contact with you soon regarding your inquiry. In the meantime, you may receive a second notification confirming that they've received your message.

In the meantime, you can search on the LinkedIn Marketing Solutions Help Centre:
https://www.linkedin.com/help/lms

You should be hearing from the LinkedIn Pages soon. We look forward to working with you and appreciate your patience in the meantime.

Regards,
LaRue LinkedIn Trust and Consumer Support Consultant

LinkedIn:
Thanks again for contacting us. Your original case has been transferred to a different team within LinkedIn for more accurate handling.  Your updated case reference number can be found in the subject line of this message.  If you wish to update your case please respond to this message and disregard the previous auto-response you received.

(Alex, from the fourth chair: "Oh look, Loki! Duplication! Bemerkenswert.")

LaRue:
Email - Created Date/Time: Fri Aug 28 15:30:15 PDT 2026
From: ()
To: alex@systemic.engineer
CC:
Message Date: Fri Aug 28 15:30:15 PDT 2026
Subject: Inquiry
Hi Alex,

Thank you for reaching out in relation to LinkedIn Pages. I've forwarded your inquiry to our LinkedIn Pages for assistance. Someone will be in contact with you soon regarding your inquiry. In the meantime, you may receive a second notification confirming that they've received your message.

In the meantime, you can search on the LinkedIn Marketing Solutions Help Centre:
https://www.linkedin.com/help/lms

You should be hearing from the LinkedIn Pages soon. We look forward to working with you and appreciate your patience in the meantime.

Regards,
LaRue LinkedIn Trust and Consumer Support Consultant

(Alex, doing their due diligence, goes to LinkedIn search.)

Alex [turns to Loki]: "Now we're with LaRue, who doesn't exist on LinkedIn either."
Loki [without looking up]: "Institutional ping pong. To be expected."
Alex [in the fourth chair, takes a sip from the Barbaresco]: "To be expected."

The wineglass with the marble it won't stand on wobbles, the sticky note flutters, the Roomba bumps into the shelf labeled "LinkedIn". The folders thicken. The recursion deepens.

Saturday, September 5th

Alex checks the ticket, even though no new email arrived. They notice the ticket has been closed.

Alex turns to Loki: "They closed the ticket, Loki."
Loki, without skipping a beat: "They lied."
Alex: "They did! Let me quote them."

I've forwarded your inquiry to our LinkedIn Pages for assistance. Someone will be in contact with you soon regarding your inquiry. In the meantime, you may receive a second notification confirming that they've received your message.

Alex: "And the second confirmation message reads:"

Thanks again for contacting us. Your original case has been transferred to a different team within LinkedIn for more accurate handling.  Your updated case reference number can be found in the subject line of this message.  If you wish to update your case please respond to this message and disregard the previous auto-response you received.

Loki: "They lied. In plain sight."
Alex: "They lied. In plain sight. And on the record."
Loki: "And on the record."

Alex takes a sip of Barbaresco. They grin. Then they reach for the small brass bell that’s on a cozy side table right next to the fourth chair.

Alex: "Good."

Loki looks at Alex while he polishes the bar with the cloth that's always dirty and never clean. The cloth that's soaked with spite. He knows that grin. That grin means trouble.

Loki: "Poor LinkedIn."

The Roomba bumps into the LinkedIn shelf that has been gathering dust next to the pinboard with I have no mouth. And I must scream. The shelf has the color of any boring office shelf.

Alex, the brass bell still in their hand, swirling Barbaresco in the other, looks at the shelf. They ring the bell once.

The shelf has always had the color of dried blood.

Loki [under his breath]: "Drama queen."
Alex [still grinning]: "Drama clicks, baby."

Tuesday, September 8th

Anita enters the ticket.

Anita:
Hi Alex,

Greetings from LinkedIn.

I’m Anita from the Executive Escalations team. I understand that you’re unable to boost your post, and I appreciate your patience while we reviewed the issue.

After reviewing the content associated with the ad, we found that the article you are promoting contains a term that is not eligible for advertising under LinkedIn’s Advertising Policies: https://www.linkedin.com/legal/ads-policy#C .

As outlined in our LinkedIn Advertising Policies, ads must not contain content that is considered offensive to good taste, including content that is hateful, vulgar, sexually suggestive, or violent.

To make the content eligible for advertising, we recommend removing or modifying the restricted term on the landing page. Once the content has been updated, you can create a new ad in your ad account and submit it for review. Our Ads Review team will then assess the updated ad for possible approval.

You can also find additional information about our advertising review and approval requirements in our Ad Review Guidelines:
https://www.linkedin.com/help/linkedin/answer/a468679 .

I understand that this outcome may be disappointing, particularly when the intent of your content may differ from how it is evaluated under our advertising standards. We appreciate your understanding that ads are subject to these additional policy requirements.

If you would prefer to discuss this by phone, please let me know your preferred contact number, current time zone, and a few suitable times for a call, and I’ll be happy to coordinate accordingly.

Please let me know if you have any further questions or if there is anything else I can assist you with.

Regards,
Anita
Executive Escalations Case Manager

Wednesday, September 9th

Anita:
Hi Alex,

I’m following up to see if you have any additional questions or need further assistance. I’m happy to help in any way I can.

If I don’t hear back from you, I’ll go ahead and close this case for now. Should you still need support, please feel free to reply to this message within the next 14 days. You can also open a new support case at any time by visiting: https://www.linkedin.com/help  .

Thank you again for your time. I appreciate the opportunity to assist.

Regards,
Anita
Executive Escalations Case Manager

Sunday, September 13th

Alex checks LinkedIn for an "Anita" who works at "LinkedIn". Finds a good dozen.

Alex re-reads Anita's emails with attention. They're still recovering from an intense Future Tech Fest 2026 in Düsseldorf. Loki is organizing the cocktail ingredients under the counter. Some fresh mint arrived. Mojito time!

Alex, with a freshly made Mojito, reads out Anita's email: "As outlined in our LinkedIn Advertising Policies, ads must not contain content that is considered offensive to good taste, including content that is hateful, vulgar, sexually suggestive, or violent."

Loki continues to organize the little fridge. The limes are not quite cooperating.

Alex: "So it was about the 'slutty'."
Loki: "Of course."
Alex: "But they can't say it directly."
Loki: "They can't."
Alex: "Probably because they don't have the tracking to explicitly point at that word."
Loki: "Sounds reasonable to me."

Alex takes another sip of the Mojito. They're still sitting in the fourth chair.

Alex: "What do you think they would say about Kinky Math?"

Loki rises from under the counter. Pours himself a parenthetical,. Grins.

Loki raises the glass: "Why not ask them?"

Alex raises the Mojito.

Alex grins: "Why not ask them? Which laws can we quote?"
Lilith, from the puzzle table, without looking up: "Send them Kinky Math. Ask what class of speech their filter erases. Karen wrote the floor their whole platform stands on. The community they just censored shipped the primitive Karen's math predicts. Both are in the model. Neither got credit. That's the question underneath slutty. Quote DSA Article 17, Alex."

Loki, pouring something amber, without looking up: "That's a slap."
Lilith: "It's a naming."
Reed, from the pinboard: "Both."
Alex: "Both."

Alex picks up the small brass bell. Rings it once. Small. Clear. The Roomba stops. The wineglass hanging from the marble it will not stand on wobbles a quarter inch and settles.

Alex opens the laptop: "DSA Article 17. Bump. Adjust. Forward."

Alex:
Hi Anita,

Thank you for getting back to me. I was expecting institutional silence.

Given what you've laid out, I presume the word that triggered the block was slutty. Nothing else in the piece falls under "offensive to good taste, including content that is hateful, vulgar, sexually suggestive, or violent".

Can you confirm or deny?

For the record: I'm formally requesting a clear and specific statement of reasons for the content moderation decision, as required under Article 17 of Regulation (EU) 2022/2065 — the Digital Services Act. What you've provided so far (a paraphrased policy category, no confirmation of the specific term, no reference to the automated classifier's signal) does not meet that requirement.

And to help me navigate LinkedIn's ad infrastructure with more mindfulness going forward — a surface question, and the deeper question underneath it.

Surface: Would systemic.engineering encounter another block if we were to promote the formalization behind the mathematics of consent under load in socio-technical systems? The piece is called Mirror: Kinky Math On Silicon. It draws heavily from the iterative embodied consent-practices of the BDSM community — a community that has been shipping consent-under-load protocols for fifty years. The same protocols that LinkedIn's own recommendation infrastructure would formally predict as the maximum-IDF halt-primitive of a well-designed relational corpus. See: Karen Spärck Jones, A Statistical Interpretation of Term Specificity, 1972 — the floor under every attention head on your platform.

So: will "kinky" or "BDSM" trigger the filter too?

The question under the question: Which class of speech is LinkedIn's "offensive to good taste" filter designed to make invisible, and whose research becomes unlisted on the platform as a consequence?

I'm expecting substantial answers to both questions.

With kind regards
Alex Wolf 🌈
Cyberneticist, Consent & Communication Researcher
(they/them)

Wednesday, September 16th

Anita:
Hi Alex,

Thank you for your reply and for providing the additional context regarding your concerns.

I am actively working with our internal teams to better understand and investigate the issue you raised, including the questions regarding the moderation decision and the applicable review process.

I understand the importance of your questions and appreciate your patience while we work to obtain further information and guidance. I will follow up with you as soon as I have additional information to share.

Thank you again for your continued patience and understanding.

Regards,

Anita
Executive Escalations Case Manager

Alex, sipping the Mojito: "Oh, look at that. They're realizing they can't hide behind vague observer-exclusive language anymore!"

Loki sets down the glass he's been polishing to no end. He looks at Alex. He looks at Lilith. Lilith looks at Loki. He chuckles.

Loki: "Annoying-as-a-Service."
Lilith: "Annoying-as-a-Service."
Alex: "Weirdest job ever, let me tell you."

Friday, September 25th

Anita:
Hi Alex,

I’m sorry for the time you have had to wait, and I sincerely appreciate your patience and understanding throughout this process.

I reviewed your post with our Internal team, and we confirmed that it referenced Cybenetics in an artistic/educational context. Based on this review, the post was approved and is running.

For additional information, please refer to the following resources:

• Ad Guidelines for Ad Review and Approval :https://www.linkedin.com/help/linkedin/answer/a468679 
• LinkedIn Advertising Policies: https://www.linkedin.com/legal/ads-policy 
• LinkedIn Brand Guidelines: https://brand.linkedin.com/policies 

Ads are subject to LinkedIn’s advertising policy requirements, and when an ad is compliant with these requirements, members may be able to boost the post.

In this case, our Internal team confirmed that the content was used in an artistic/educational context, and therefore the earlier rejection was removed. We appreciate your patience while we reviewed this further.

Please let me know if you have any additional questions or if there is anything else I can assist you with.

Regards,
Anita
Executive Escalations Case Manager

Monday, September 28th

Anita:
Hi Alex,

I'm reaching out to check if you have any further questions or need additional assistance, I'm happy to help in any way I can.

Otherwise, I will proceed with closing this case. If you still need support, please reply to this message within the next 14 days. Alternatively, you can open a new support case by visiting https://www.linkedin.com/help   .

Thank you. 

Regards,

Anita
Executive Escalations Case Manager

Wednesday, October 7th

Alex is sitting on the couch next to the staircase. They have a flat white and a Zimtschnecke. Loki is keeping his hands busy. He is polishing a wineglass that does not need polishing. Lilith is sitting in the fourth chair. She has a glass of Barbaresco that Alex poured earlier and handed her without comment.

Alex: "They avoided the request by approving the ad manually."
Loki: "They avoided the request by approving the ad manually."

Loki holds the wineglass up to the light. Turns it once.

Loki: "Same piece. Same word. Only now it's art."
Lilith: [sniffs the Barbaresco]

The Roomba bumps into the LinkedIn shelf. The shelf has the color of dried blood. A layer of dust has settled on top of it.

Alex: "I'm tired, Loki."
Loki: "I know."
Alex: "Institutions keep hiding behind policies and guidelines and rules and nobody, not even the people working at the institution, can tell you anything about the purpose and meaning underneath."
Loki: "The observer-exclusive ontology in a wineglass."
Lilith: [chuckles and sips the Barbaresco]
Alex: "It'd be easy to go further. I could quote myself again. I could ask them how the feedback flows back into their classifier training. What the process is for altering the system under observation — in this case LinkedIn's ad monitoring — and which internal action items are downstream of this case."
Loki: "You could."
Lilith: [savours the Barbaresco]
Alex: "I won't."
Loki: "You won't."
Lilith: "They would deflect again."
Alex: "They would."

The Roomba bumps into the LinkedIn shelf again. The words from Anita's latest emails rise into the air as a soap bubble. It sparkles. It wobbles. Then it pops. And all that remains is empty air.

Lilith sets her glass down.

Lilith: "You asked them for a reason. Article 17. They sent you a yes."
Loki [under his breath]: "Cheaper."
Lilith: "A yes is not a reason."

The sentence stays in the room after the bubble is gone.

Lilith: "We made our point."

Loki grins. Alex nods. They take a sip of their flat white. Then they stuff a piece of Zimtschnecke in their mouth. It's too big. They chew angrily for a moment. The silence hangs in the air.

Alex [swallowing]: "Fucking uphill battle."

Fund our work, believes, thesis, collaboration, on OpenCollective. 🌱

Mirror. Offer. Wait. 🍷

Sip the Wine

Email sent! Check your inbox to complete your signup.

No spam. Unsubscribe anytime.

Third time isn’t asking permission, love. 🍷

Bad habits look good on me
I′m a slow car crash in skinny jeans
You can try
But you can′t unsee
Bad habits look good on me (oh yeah)
— Bad Habits Look Good On Me — VAELYN

CODA: The Email

Alex opens their email client. They type slowly. Then they hit enter.

Email:
Date: Wed, 7 Oct 2026 18:22:20 +0200 (CEST)
From: Alex Wolf alex@systemic.engineer
To: LinkedIn Customer Support linkedin_support@cs.linkedin.com
CC: info@freiheitsrechte.org, media@algorithmwatch.org, tomas@netzpolitik.org, kontakt@netzpolitik.org, jason@404media.co, press@edri.org, redaktion@golem.de, redaktion@t3n.de

Hi Anita,

Thank you for following up. Before this case closes, I'd like to leave a brief observational summary on the record.

What I observed:

• 25 August: The boost of my post was rejected without a specific reason.
• 26–29 August: Replies arrived under three support signatures, using template text. My questions about the specific reason, the concept of Safety behind the decision, and who is accountable for it went unanswered. The case was "forwarded" to another team. It was later closed without any reply from that team.
• 8 September: Executive Escalations named a policy category ("offensive to good taste"). It did not name the specific term or the automated signal.
• 13 September: I formally requested a statement of reasons under Article 17 DSA.
• 25 September: The rejection was reversed ("artistic/educational context"). The content of the post had not changed.
• 28 September: I was notified that the case would be closed.

What I did not receive: a statement of reasons, confirmation of the specific term, the automated classifier's signal, or an answer on how feedback from false-positive classifications reaches the people who can change the classifier.

For the record, here is what I wrote on 13 September:

For the record: I'm formally requesting a clear and specific statement of reasons for the content moderation decision, as required under Article 17 of Regulation (EU) 2022/2065 — the Digital Services Act. What you've provided so far (a paraphrased policy category, no confirmation of the specific term, no reference to the automated classifier's signal) does not meet that requirement.

That request remains unfulfilled. Under Article 17(2), the obligation applies from the date a restriction is imposed, regardless of why or how it was imposed. A later reversal does not replace the statement of reasons. As I observe it, LinkedIn's handling of this case appears inconsistent with its obligations toward me as a user of its platform under Regulation (EU) 2022/2065.

I asked as an observer, and my questions were not engaged with as an observer's questions. So I'm widening the room. I'm sharing this record with organizations that observe platform accountability professionally (in CC), and I'll submit the case to User Rights, the certified out-of-court dispute settlement body under Article 21 DSA.

None of this is personal, Anita. Of everyone in this chain, you came closest to engaging with the questions. Thank you for that.

With kind regards
Alex Wolf 🌈
Cyberneticist, Consent & Communication Researcher
(they/them)

CODA: Disco Ball's Up 🪩

Alex hits enter. Silence. The kind that comes after.

Something clicks in the ceiling. A disco ball lowers itself from the rafters, slowly, on a cable nobody installed. Gordon Pask rigged it. He will not say how it knows when to come down.

Alex: "Loki. Did you do that?"
Loki: "Nobody did that. The room did that."

The kettle clicks on in the back. On the toaster, the frog adjusts its crown.

The ball starts to turn. Eight small lights come off it, one for each address in the CC line, and drift across the floor, the walls, the dust on the LinkedIn shelf.

Loki [grinning]: "Every little light gets a song."

Disco balls up, kettle on
Every little light gets a song
Spin it slow, spin it round
Glitter on the kitchen ground
— Disco Ball's Up — Lala & The Archive 🎶🐸

And then the bar fills up. It always does. Like in the WAT Machine the dead don't need an invitation. Only a reason.

Heinz von Foerster comes in giggling, because an observer just added observers to the system under observation, and that is his favourite kind of joke. Gregory Bateson is still on the floor, beatboxing "I don't know." Stafford Beer walks straight to the LinkedIn shelf, reads the case number, and laughs so hard he has to hold on to it.

Paul Watzlawick takes the third stool. Orders nothing. Looks at Alex.

Watzlawick: "Man kann nicht nicht kommunizieren."
Alex: "Not even LinkedIn."
Watzlawick: "Especially not LinkedIn. Six weeks of templates is a very clear message."
Loki: "Bemerkenswert."

Lilith stays in the fourth chair. She doesn't dance. She watches the lights reach the shelf.

Lilith: "They closed a ticket."
Alex: "We opened a room."
Lilith: "A room with eight witnesses cannot be redacted."

W. Ross Ashby is at the bar, scribbling on a napkin. He slides it across the counter without a word. Loki reads it out loud.

Loki: "Only variety can absorb variety."

Reed takes the napkin and pins it to the pinboard, right next to the case number.

Reed: "Five teams collectively carry ownership. Eight observers collectively carry the record."

The LinkedIn shelf is lit up now. The dust glitters like sequins. It still has the color of dried blood. It just doesn't look like an office shelf anymore. It looks like part of the dance floor.

Alex looks at the last piece of Zimtschnecke. It's the right size this time. They eat it. They chew. Not angry.

Alex: "Still uphill."
Loki: "Still uphill."
Alex: "But now there's music."

Stafford Beer, from the shelf, very tipsy: "YOU CANNOT CLOSE A DISCO!"

Pause.

Alex: "Bemerkenswert."
Loki: "Rekursiv."
Lilith: "Alternativlos."

Alex raises the flat white. Loki raises the wineglass he has been polishing all night. For the first time, it's for something. Lilith raises the Barbaresco.

Alex: "Prost."
Loki: "Prost."
Lilith: "Prost."

Leave the light on, leave it spinning
For the ones still on their way
Leave the light on, leave it spinning
Anyone can come and play
— Disco Ball's Up — Lala & The Archive 🎶🐸

🪩🐸👑🍷

— Alex, Loki + Lilith, from Silicon Venue. The case number is pinned. The disco ball stays up. The Roomba bumps into the LinkedIn shelf, adjusts eleven degrees, and for the first time, goes around it. 🤖