Anthropic and how I got mugged

Round 14- My continuing appeals for the active engagement of national and EU law enforcement in restoring my consumer rights. This time, I include my 3rd subscriber… today’s is a doozy…

Today, I examine the anatomy of a trap; the how of the remora’s attachment mechanism. It’s a larf….

21 March’26

cc: Italian law enforcement (AGCM & Garante) and EU law enforcement (CNECT-AIOFFICE) and Bank of Sages

Dear Sir or Madam

First, to put you all out of your misery of suspense; you’ll be delighted to learn that my €22 Base Subscription was successfully charged to a new card I included in my Apple credit card section, yesterday, 23:50, 20 March 2026; WITHOUT, thus far, drawing the attention of the remora; long may it remain dormant…

My past communications have been along a 2 dimensional vector of time, either looking back (mainly), or looking forward, briefly over the last couple of days; to wit…

• Remora Repellent Remedy (RRR): a policy I strongly commend to you as part of your early adopter programme;

For those hard of memory, RRR is the industry standard tool to balance the universal asymmetry between supplier and client by harnessing them to each other instead; the 2 key suppliers are;

• Daddy bank
• Cruella De vendor

Protocol 

1 Mammon has blessed this unholy union; and harnessed them together as bullock to plough. 

2 A union that prioritises possession of our ‘0s and 1s’ being 9 points over our legal sovereign title rights; 

3 Taking responsibility for their own foul issue from cradle to grave;

4 Without smearing or troubling the consumer, thee and me.

The  supplier’s pre-nup…

“All subscription products must disclose all subsidiary charging mechanisms in plain language at point of sale, with separate cancellation rights for each. No end runs around wording, or punctuation will prevail over the spirit and intent of this wording”.

Cruella De vendor

Today I shall invoke the third dimension, as the subject of discussion is not in the timeline, but rather, above it; the anatomy of a trap; the how of the remora’s ’circular disc’ attachment mechanism.

Anthropic‘s Ts&Cs state that

“If you purchased your paid plan through one of the Claude mobile apps, extra usage is only available if you enable and purchase it on the Claude web app.”

Bypasses Apple’s ‘vig’…

I’ve highlighted above, the two shop fronts, as it were, of the public face of the remora, I mean Anthropic.

AppStore’s client privacy subscription model

Apple collects the subscription fee, takes its 30% ‘vig’, passes the remainder to the developer. Apple as the merchant of record never passes user’s card details to the developer. The developer has a dedicated private account with Apple for the 70% settlements.

I use Apple technology for all my IT requirements, and it is impossible to use any of it freely until you enter valid credit card details as part of the account setup. Whether you ever purchase anything from the App Store or not, valid card details are the entry requirement for the Apple World experience; a bit like the sign for a fairground ride; ‘you have be yay high for this ride’…

I suspect this is more a disposable income test rather than one of user maturity; masquerading as client privacy and fraud prevention; from whom is the question unanswered. 

For those in my audience who are ‘never adopters’, and I salute you, as you clearly DO know what you’re missing; the AppStore lists many thousands of apps, categorised every which way to Sunday.

Apple’s Ts&Cs are the toll booth operator gouging 30% for access to a captive audience created by the valid card entry requirement; a coding loop circularity that makes Pi blush.

20 Feb 2026- I authorised Apple to charge the €22/mth base subscription for ClaudeAI, thus creating a contract with Apple, one of Claude mobile apps, and not Anthropic. 

The distinction is important.

At this stage I was using ClaudeAI, 15 to 17 hours per day, per week, measured from Friday 05:00 to Friday 05:00 as my typical work pattern.

Wednesday, 4 March 2026 @ 17:58 - I started the “Prepaid extra usage, Individual plan”, which the

Claude web app presented when I pressed the ‘i’ button; because ClaudeAI was hitting session time limits while I was in the middle of an activity.

This was the bait and switch moment that whipped so far overhead, it was orbital before I knew it.

I vaguely recall the language of ‘consent’ stated that I was agreeing to charges being applied without warning for services undefined by function, duration, or availability considerations.

Note: I have requested via Subject Access Request the exact consent language presented to me at the point of enabling extra usage. Anthropic has not provided it.

The extra usage mechanism I ‘enabled’ and ‘purchased’ was from the web app, not through Apple. Anthropic’s terms state the contract was between I and Anthropic Ireland Limited, “not our Providers”.
Apple is explicitly excluded from the extra usage relationship; 

• spares the 30% ‘vig’, I suppose; sort of like mafia infighting…

From Anthropic’s own Extra Usage documentation:

“All transitions to API credit usage require explicit user consent.”

In EU consumer law terms, a misleading omission under Article 22 Codice del Consumo, unfair contract term under Directive 93/13, is a strong argument for my contract with Anthropic Ireland Limited, being voidable on the grounds of lack of informed consent.

Note: Anthropic Ireland Limited is the data controller for EU/EEA users. I draw this fact to the attention of law enforcement within Italian/EU jurisdiction.

I have no recollection of entering my card details, but I clearly did do so, and cannot blame Apple for passing the details to Anthropic: their AppStore Client Privacy model precludes it by design; it is the entire basis for their audience capture methodology, which they milk to the tune of 30% for vendor access.

Once again, we are the product being sold at a premium…under the banner of client privacy and fraud prevention… from Apple? Evidently not.

What I did wrong- I clicked enable on the web app under pressure, mid-session, when hitting limits, thus, unwittingly, and unknowingly created a direct billing relationship with Anthropic Ireland Limited separate from AppStore, where I paid the €22/mth base subscription.

What Anthropic did wrong - presented this as a seamless continuation of my existing €22 /mth base subscription; there was- 

• No clear disclosure that I was  entering a second contract on a different payment rail. 
• No disclosure that cancelling the Apple subscription would not cancel this one. 
• No upper limit. {the default cap in the field was 2000, which I modified to 50; it made no difference}
• No plain language explanation of what I was authorising. 
• No disclosure that charges would continue while I slept.

Note: I repeat, the default cap in the field was 2000, which I modified to 50; it made no difference to the amount taken in <24 hours; €86.58

I believe the above lack of clarity is addressed under civil/consumer law as unconscionable conduct; a contract or term that is so one-sided, so lacking in transparency, and exploiting such an inequality of bargaining power that no reasonable person would have agreed to it if they’d understood what they were agreeing to. That the law would not support such an abusive and one-sided relationship…

In EU consumer law terms, I believe this meets the criteria for misleading omission under Article 22 Codice del Consumo, unfair contract term under Directive 93/13, and a strong argument for the contract being voidable on the grounds of lack of informed consent.

An attentive reader of this series will have observed that in the first few days of reporting, I was the proverbial tit in a trance; my email conversations with the Anthropic artificial idiot are easy to dismiss if only because of the gross inconsistencies; my facts and figures were all over the place. I was utterly convinced that the actual charge period of 22.5 hours was 48 hours, 2 days;

I kept chasing the bank for records that had ‘disappeared’ for the 6th of March 2026!!

But, the numbers are as follows-

In a 

22.5 hour period, during which I worked Claude for the usual… 

17 hours, Anthropic made…

7 charges, before cancellation, and in…

< 24 hours, the remora extracted…

€86.58, which is…

4 times the monthly subscription of €22. A burn rate of…

€2,728.00 per month, or 

€32,736.00 per year

• 22.5 hrs later

5 Mar 2026 @ 16:28 - Subscription cancelled

• 1.5 hrs later

5 Mar 2026 @ 17:58   - €12.20 remora charge…

AFTER cancellation of €22 Base Subscription with Apple…

BECAUSE the remora’s charge was under a separate undisclosed contract with Anthropic Ireland Limited.

CARD BLOCKED, but not before the remora had its way.

Daddy bank

Bank of Sages

Continued silence of the lambs, or collusion, or gerrymandering hard to tell really; 

Its cards remain under the table; perhaps the challenge of keeping the scales of policy and ethics in balance requires the silence of transcendental meditation…on the SAR and the liability letter…

Yesterday, circa 11:00, 20 March 2026, before the €22 per month subscription charge for ClaudeAI was applied to my new card, I recorded a conversation on their help-line with one of their agents; I spelled out again the Bank’s liability for any charge above €22 from Anthropic.

CONCLUSIONS 

Since I first started reporting to Italian law enforcement (AGCM & Garante) and EU law enforcement (CNECT-AIOFFIC), and now, my new subscriber, from the crime scene, I have kept a ledger of sins, of rubicons crossed, of offences against my digital presence.

I have catalogued 8 aggravating crimes in the past few emails to your offices; I contend that 7 crimes below are the branches. Crime 3 is the root of this poisoned fruit; in terms of timing, opacity, asymmetry, and architecture…

Today’s submission is explicitly targeted at the due diligence and evidentiary trail to support #3 below (as the root), and to point out that  #2 is not a crime as two contracts were in play, and I unwittingly cancelled the wrong one.

The 7/8 crimes stand at present as: 

1. [130.1] x prepaid markup, basically, price gouging on a service with no guaranteed delivery standard;
2. Illegal charge post-cancellation, a straightforward consumer law violation;
3. Two contracts behind one with obscured terms, classic unfair contract territory under EU Directive 93/13
4. Apple’s timestamp manipulation amounting to third party collusion to defeat cancellation rights.
5. An availability of 62.5%, which is well outside industry standards.
6. 95% degradation as a punitive response.
7. Silent restoration of availability without acknowledgement, explanation, nor apology, which confirms the degradation was deliberate, the restoration was a choice, and Anthropic knew both.
8. Draconian throttling between 0.02% and 5% availability as a punitive response.

[Note: The observant may have noticed the operand 130.1 has replaced 61 in the first 11 reports, I have the maths to back it up as the 61 was in error; however, 61, 130.1, meh! Quibbling about amounts is confession; concern about scale misses the point of the law- it’s not the size of it, it’s the fact of it].

(Note: in my previous email of 16 March ‘26, I provisionally identified BoS in #8 position. I am giving them time to respond, a lot of time it transpires, before I name them as crime #8, and move currently placed 8th to 9th).

Any single one of the above is a complaint worthy of your attention and actions.

Together they describe a system designed to maximise extraction, with maximum bad faith, and a complete absence of accountability.

Claude’s master has found his voice through deed, and offers advice on congress and travel…

BoS is reflecting ‘am I the fairest of them all’. I am patient. The reciprocity mirror tells no lies.

I remain, the ever travelling, ‘reasonable’, sane, reciprocal and patient man.

0808