Bonum Certa Men Certa

European Patent Office (EPO) Illegally Turning to Slop Behind Closed Doors, Staff Objects to This Hidden Catastrophe

posted by Roy Schestowitz on Mar 25, 2025,
updated Mar 25, 2025

This is what happens when Europe's second-largest institution gets stacked or stuffed with white-collar criminals who lack any grasp of science and lack respect for truth/facts

Slop - or sometimes LLM slop (basically nonsense spewed out by models that lack actual comprehension of both inputs and outputs) - is practically worthless hype and a bubble, just like many that came before it and continue to re-emerge [1, 2] to reinvigorate and deceive. Remember 4IR? Or IOT/IoT? How about "Metaverse"?

Nowadays almost everything that MBAs touch is being re-labeled "hey hi" (AI). It's not new, so they use terms and redundant adjectives like "advanced" or "generative" or "agentic" or "superintelligent" "AI". They have marketing slogans with terms like "era" or "arms race". There's hardly any use case for it... to date, except gimmicks or harder-to-detect plagiarism. To give a new example of what we're dealing with, yesterday we said that Planet Ubuntu (or Ubuntu Planet) is LLM slop because it aggregates Serial Sloppers or bots. Today too we saw another example of it, again with the same culprit in Planet Ubuntu (or Ubuntu Planet):

Linux Fixed Release vs. Rolling Release Distributions: Which One is Right for You?

Slop as usual.

Linux distributions come in two main release models: fixed release and rolling release. Each has its own advantages and drawbacks

Now, imagine dealing with serious legal issues (with a lot of money at stake) using slop. What sane person would do this? Probably these corrupt people who do illegal things behind back doors and closed doors. They conspire to break the law and/or harm people. Such people must never be in positions of power at all.

Thankfully, the illegal EPO policy ("Internal use only") has been leaked:

EPO AI POLICY: It's not legal, keep quiet about it

Hiding one's illegalities behind a veil of secrecy might be expected from rogue and illegal (unconstitutional) pseudo-agencies like "DOGE" [sic]. The EPO is no better, except this is happening in Europe and it almost directly discredits the EU.

Thankfully the EPO's Central Staff Committee has already analysed it and politely explained what's happening here. To quote:

Dear Colleagues,

Staff of the EPO and their representation discovered in the news item of 5 March 2025 that the EPO has adopted an AI policy. The document was adopted without consultation of any stakeholder and has far-reaching consequences which should have been subject to an open debate.

First, the EPO AI policy is not bound by international texts such as the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy, and the Rule of Law of 17 May 2024, and the EU AI Act.

Second, the EPO AI policy lacks boundary between AI "assistance" and "decision-making". While the policy presents as key principle a “human-centric approach” promising that “final decisions will be taken by humans at the EPO”, the document nevertheless advocates the implementation of AI in pre-search, machine translation, pre-classification, re-classification and “to potentially automate considerable parts of the search” [and] "to automate repetitive tasks, leading to increased efficiency and productivity in various areas ? including the examination process”.

In the already implemented AI-based automatic closing of circulation of patent applications for the purpose of classification, the classification decision is done without the final input of the relevant classifiers and gérants in the field who are simply supplanted by AI.

Third, the EPO AI policy acknowledges that “the automation of tasks through AI may lead to job displacement and the need for reskilling programmes to support affected individuals” and hence directly acknowledges an impact on working conditions. Further far-reaching changes on adverse decision-making on employee rights and benefits are foreseen via AI systems which are classified as high-risk.

In this open letter, we request that the present EPO AI policy is submitted to an open and transparent debate with all stakeholders and to statutory consultation.

Here's the full open letter as GemText, HTML, and plain text:

European Patent Office
80298 Munich
Germany

Central Staff Committee
Comité central du personnel
Zentraler Personalausschuss

centralSTCOM@epo.org

Reference: sc25022cl

Date: 21/03/2025

European Patent Office | 80298 MUNICH | GERMANY

To:
Mr António Campinos
(President of the EPO)

By email:
president@epo.org

OPEN LETTER

EPO AI policy: A policy requiring open debate and consultation of stakeholders

Dear Mr President,

Staff of the EPO and their representation discovered in the news item1 of 5 March 2025 that the EPO has adopted an AI policy2. The document was adopted without consultation of the staff representation, without any decision from the Administration Council and is classified for “internal use only”, thereby suggesting that there is no intent to make it available to the public. This policy has far-reaching consequences which should have been subject to an open debate.

The EPO AI policy is not bound by international texts
First, the policy states that the EPO just takes note of initiatives such as the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy, and the Rule of Law of 17 May 2024, and the EU AI Act. However, further into the document, the EPO hides behind its administrative autonomy under Article 4(1) EPC to consider these texts as not legally binding3. We note that the same approach was adopted by the EPO to argue that it was not bound by the European Convention on Human Rights. This led our Organisation to several breaches of the right to freedom of association (incl. the right to strike

_______

1 “New EPO AI Policy adopted”, News item of 5 March 2025

2 EPO AI Policy: Policy on the use of artificial intelligence systems

3 page 5, par. b


and the right to freedom of communication). We believe our organisation should take utmost care to learn from the past and explicitly recognise international texts, and in particular when it comes to AI.4

Lack of boundary between “assistance” and “decision-making”
Second, the policy presents as key principle a “human-centric approach”5 defined as “the combination of human + AI providing better result in terms of quality and efficiency than either of them alone” while “final decisions will be taken by humans at the EPO”. The document also recognises that AI “may even supplant decision-making processes”6 and “[i]n public administration this may pose certain legal questions relating to due process”. However, the document advocates the implementation of AI in pre-search, machine translation, pre-classification, re-classification7 and “to potentially automate considerable parts of the search” [and] "to automate repetitive tasks, leading to increased efficiency and productivity in various areas ‒ including the examination process”8.

There is already an area where AI has started to supplant EPO officials and that is for the classification of patent applications. The reason provided is consistency9 : “AI can ensure consistent practice and administration, e.g. by categorising applications into relevant classes and subclasses, improving the accuracy and consistency of patent classification.” In practice since June 2024, the EPO has put in place an AI-based automatic closing of circulation of patent applications. Allocated classes by third-parties can be marked by the AI as final without any confirmation from classifiers and gérants in the relevant fields. The classification decision is done without the final input of the relevant experts who are simply supplanted by AI but remain, according to this policy10, subject to “principles and rules relating to the responsibility and accountability of EPO employees laid down in the Service Regulations and other internal rules”. The feedback we gathered from classifiers and gérants is for the vast majority negative about the quality of the AI-based automated classification when currently 46.7% of the classification tasks are automated. We are deeply concerned that the Office wants to continue to leverage AI in classification under these conditions and reach by 202811 a target proportion of 90% of classification tasks that no longer require human intellectual classification.

______

4 page 5, paragraph b

5 page 2, last paragraph

6 page 6, section C. 1. paragraph 1

7 page 3, paragraph 2

8 page 3, section 2. bullets 1 and 5 bullet

9 page 3, section 2. bullet 4

10 page 9, section 4. paragraph 3

11 Strategic Plan 2028 Progress Dashboard


Impact on working conditions
Third, the policy acknowledges that “the automation of tasks through AI may lead to job displacement and the need for reskilling programmes to support affected individuals” and hence directly acknowledges an impact on working conditions. Further far-reaching changes on adverse decision-making on employee rights and benefits are foreseen via AI systems which are classified as high-risk. They are defined as such if they are intended to be used12:

• “to evaluate the eligibility of individuals for essential benefits and services, including coverage under healthcare insurance provided or paid by the EPO, as well as to grant, reduce, revoke, or reclaim such benefits”

• “for assessing the likelihood that an individual will engage in misconduct not solely based on the profiling of individuals as referred to in Article 3(1)(d) EPO DPR or to assess personality traits and characteristics or past behaviour of individuals or groups”

• “for the profiling of individuals as referred to in Article 3(1)(d) EPO DPR in the course of the detection or investigation of misconduct”

While such AI systems are classified as high-risk, their implementation or use is nevertheless considered if individual and overall residual risks “are deemed acceptable” and provided that AI is used for its intended purpose and with the promise that they will be trained, tested and validated with the “data sets that meet the highest quality criteria and are subject to appropriate data governance and management practices”13. This is clearly insufficient to reassure staff.

The EPO AI policy has triggered many discussions among concerned EPO staff.

We request that the present EPO AI policy is submitted to an open and transparent debate with all stakeholders and to statutory consultation.

Sincerely yours,

Derek Kelly
Chairman of the Central Staff Committee

______

12 page 12

13 page 13

The only room/capacity for slop at the EPO should be none, zero. People don't pay for bots to pretend to do work, the EPC didn't see member states reconvening on the matter.

Benoît Battistelli used to ride all sort of buzzwords (vague mumbo-jumbo for voodoo science) to justify granting loads of European software patents - i.e. patents which are both illegal and undesirable. António Campinos has done the same with "blockchain" hype and now "hey hi". Since Campinos has no background in science, name-dropping buzzwords is the best he can ever do. We did some video rebuttals to his laughable name-dropping. Is this clown the best Europe has to offer in its biggest patent office? Highly concerning!

The EPO's crises deepens the financial crisis in Europe and contaminates the EU with corruption. Who stands to gain from all this and at whose expense? █

Other Recent Techrights' Posts

An American and South Korean Cartel Artificially Inflating the Prices You Pay for Computers (or Any Service That Relies on Computers, I.e. Almost Everything)
As long as the principal tenant of the White House is a white-collar criminal we can expect only cover-up and participation, not a crackdown
The EPO Still on Strike. 141 Tickets Opened Regarding Lack of Toilet Paper (or Similar).
Yesterday staff representatives at the EPO in The Hague (LSCTH] spoke of a meeting which took place a month ago
 
Intercontinental Brigading - Part XXXXXIV - Hostility Towards Media, Coming From People Who Promote Microsoft Back Doors and Engage in Mass Plagiarism, Attack Women Etc.
The core issue or the seminal issue is Microsoft as the "master" of all computers
Trading 212 UK Limited Acting Like a Scam, FCA Says It's Legitimate
If this was some coy prank or a scam, it is certainly not funny
Richard Stallman's Talk Earlier Today and Next Month's Appearance in Mexico
He seems to have managed to dodge the 'cancel mob' for quite some time now
The Cost of Losing Integrity at The Register MS
Recently we learned that they make a "ton of money" out of their Huawei spam
Links 17/11/2026: Jensen Huang's Price-Fixing Story, "Russian Attacks Aimed At 'Breaking The Spirit' Of Ukrainians"
Links for the day
NVIDIA Continues Diving and Press Reports Speak of Business Crimes of Jensen Huang (Not China or Sex)
The deeper they go...
Gemini Links 17/11/2026: Super Power, Knowember, Writing in Shorthand, Cooking over a Woodstove, Decoupling Code and Content
Links for the day
Links 17/11/2026: California Sues Slop Peddlers and Nvidia Was Already in Trouble Before the Jensen Huang Mess
Links for the day
NVIDIA Corp Was Grossly Overvalued, Lost 36.06% of Alleged "Market Value" in 5-6 Weeks
NVIDIA has many of "obligations" (to pay), not money
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Monday, November 16, 2026
IRC logs for Monday, November 16, 2026
Intercontinental Brigading - Part XXXXXIII - An Attack on the Dignity of Reporters and Their Family Members (by Proxy/Attack Dogs)
The Microsofters seem to be super-upset that I uncovered Microsoft fraud
IBM and OSI Said opensource.net Would Replace opensource.com, But It Has Been Dead for 1.5 Years and the Person Who Had Made Promises Left OSI
Over a year ago he left
Red Hat and IBM Meet Targets (of Pushing People Out)
now there are claims of security breach inside IBM, but no press reports related to that
Gemini Links 17/11/2026: Reading Clocks and StreetComplete
Links for the day
Buy Influence "Over Dinner With The Register."
The Register MS is the facilitator or 'pimp'
While the Slop Industry Wobbles The Register MS Bags Tens of Thousands of Pounds to Publish Propaganda With "AI" 23 Times in It
The Register MS is selling its reputation away
The Real News About the White House (Not the 'Exclusive' Rupert Murdoch Published for It at Midnight Monday) is Complicity in NVIDIA Fraud
They need a distraction from commercial malpractice
Links 16/11/2026: Microsoft Increases Indian Clown Computing Prices by 15%, Meta Fined $887 Million in Italy
Links for the day
Gemini Links 16/11/2026: Container Gardening, Laughing at an LLM, "EU 72-Hour Guidelines Through a Lawyer's Eyes"
Links for the day
I Tried Linux Again After 10 Years - Here's What Blew My Mind
It's like leaving your hometown as a teenager and coming back to find the weird kid from high school is now the mayor
Links 16/11/2026: Data Centers (White Elephants) as Tax Evasion Scam/Ploy, Social Control Media Harms Boomers, "Microsoft Outage Hindered Ticket Issuance in Greece"
Links for the day
The Register MS Seems to Have Moved to 'The Cloud' (WeWork) After Shutting Down the Office
The Register MS will have a hard time explaining why it so passionately participated in the "hey hi" pyramid scheme
Holiday Plans (We Need More IBM Whistleblowers)
in particular with possession of material in relation to PIPs, NDAs, and silent RAs
After Microsoft Outage GNU/Linux Usage Leaps to 9% in the UK (Where Public Services Were Notoriously Overdependent on Microsoft Disservices)
in the UK we are extremely dependent on Microsoft stuff in the NHS, courts, law firms, police etc.
Richard Stallman Was Right About the Corruption of the Term "Artificial intelligence"
This will do severe damage to the administration which participated in the scam
The Dictatorship Already Working to Distract From NVIDIA Corruption and Slop Bubble Reaching Its Limits
Circular financing is simply a form of financial fraud
IBM Selling Itself, Piecewise, to Raise Money (to Pay Salaries)
IBM is a zombie waiting to inhale one last time (last breath)
"Allies" of Disgraced NVIDIA CEO Aren't About Morality, They're Just Covering Up Their Own Backs (and Loans)
The next stage is examining the false accounting (fraud basically)
No Outage Since 2023
I did host from home between 2020 and 2023
Intercontinental Brigading - Part XXXXXII - Back Doors and Kill Switches From Microsoft Are Objectively Undesirable
"I dual-boot Omarchy and Windows. Be careful updating your BIOS. Mine made Omarchy unbootable."
Gemini Links 16/11/2026: Enjoyable Coding, Social Control Media Considered Harmful, and Recursive Rectangular Subdivision
Links for the day
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Sunday, November 15, 2026
IRC logs for Sunday, November 15, 2026
Cusp of Collapse
NVIDIA Corp (NASDAQ: NVDA) is set to open in the red tomorrow
Getting Fed Up Not Only With Microsoft But Also Windows
This morning we wrote about North America
Layoffs and Silent Layoffs (e.g. PIPs) as a "Break" From IBM
If the press doesn't report it, then it never happened, right?
Microsoft XBox at 25 and All-Time Low
XBox will be "officially" dead (not the brand) when sales fall below some threshold and then production is halted, sort of like Apple's Vision Pro
Links 15/11/2026: Microsoft Already Emitting Chaff to Distract From Its Catastrophe, Jensen Huang Deemed Just 'Fall Guy' Amid Collapse of the Pyramid Scheme
Links for the day
Gemini Links 15/11/2026: CD Players Live on and Self-Hosting from Home Fallacy
Links for the day
When Short is Way Too Short (Superficiality, Lowering of Entropy)
Demonstrating how truly worthless social control media was all along
Techrights Hates Slop Bots
They benefit nobody
The Economy is Fine! (If Only for China!)
The future is slop! If not, then it's China's fault! It's "sabotaging" the "Western hey hi" in this "hey hi arms race!"
"IT" as "Musical Chairs" (on the Deck of The Titanic)
Each generation will get worse for "IT"
Salesforce in Trouble, Debt Nearly Quadrupled in 9 Months
Salesforce is perishing in a deep pool of debt
"Microsoft is Next IBM"
The words of a Microsoft insider and responses to that
IRC is Decentralised
Bye, Microsoft Teams/Skype
Links 15/11/2026: Murdoch's "WSJ Buries Stephen Miller's Racism" and "America Has Entered a New Age of Propaganda"
Links for the day
In the UK, GNU/Linux Usage Soared During Microsoft Downtime
since the massive outage GNU/Linux usage soared to over 10%
Google Slop About Richard Stallman Shows Stallman Was Right, LLMs are BS Generators
Google slop is the "new Google"
Spouse of Garrett Needs to Get "Several Firearms" to Feel Safe Amid Divorce (Garrett Has a "Girlfriend" Now)
Seems like Isis and my wife have a lot in common when it comes to Garrett
Vista 11 is Small
Can Microsoft survive the fall of Windows?
Six Richard Stallman and FSF Talks This Month (FSF Staff at Ohio State University)
Dr. Stallman's talks in Ecuador are well timed because of the collapse of the LLMs (slop) pyramid scheme
NVIDIA's Crisis Has Only Just began
NVIDIA Corp (NASDAQ: NVDA) is set to open in the red tomorrow
IBM Has Destroyed Red Hat and Hashicorp, Insiders Say It's the Usual From IBM:
Bluewashing is a death wave
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Saturday, November 14, 2026
IRC logs for Saturday, November 14, 2026
Techrights Was Right About Slop Destroying Development Communities and Projects
Slop does not improve your project. It kills the community. Without community, what is left? A 'botfarm'?
Intercontinental Brigading - Part XXXXXI - Harassment Law and "Things Are Unfairly Asymmetric" (Exploring the Boundaries of Death Threats and Various Threats From America)
the sort of "technofascism" that our country must learn to confront rather than facilitate
Gemini Links 15/11/2026: "The Bastards Can Never Win" and "The Last Low-Tech Generation"
Links for the day