Bonum Certa Men Certa

Global Patent Warming

I cannot grant those patents. So I will change the rules.



Summary: The old term "Global Patent Warming" comes to mind when one assesses the neoliberal approach of today's EPO, where the sole goal is making piles of money by granting loads and loads of illegal European Patents

UNDER the reckless management of Campinos and Battistelli the European Patent Office (EPO) makes it a lot easier for examiners to allow European software patents (sometimes compelling them to grant in defiance of the EPC because of the misguided guidelines), at least within the Office, not outside (national courts). Lawyers would admit and they occasionally say that it's even harder to get software patents from the USPTO (after AIA and 35 U.S.C. €§ 101) than from the EPO.



"Off-the-shelf Free/libre software libraries allow programmers to paint just about everything as "hey hi" within less than one hour."It's very troubling to us, especially geeks. The EPO makes it easier to get illegal patents on statistics and mathematics by misusing buzzwords like "hey hi" (AI) -- a term that nowadays refers to all sorts of things ranging from automation to computing. Off-the-shelf Free/libre software libraries allow programmers to paint just about everything as "hey hi" within less than one hour. It's very 'plug-n-play'; and still... it all boils down to algorithms.

Grzegorz Wesela-Bauman (JWP Patent & Trademark Attorneys) has published in Mondaq this piece entitled "New EPO Guidelines – Easier Procedures For Patenting AI-Based Inventions" and we'll quote just the relevant pieces of text while highlighting key parts:

The EPO has made considerable changes to both the procedural issues, which are important for patent attorneys, and in recognizing the patentability of inventions, which is of importance for inventors.

Among the procedural changes, it is worth pointing out the simplified representation for applicants before the EPO before granting the patent, or the simplified method of obtaining discounts for application payments and substantiative examinations when there is more than one applicant.

Substantive changes include granting AI-based inventions the status of technical solutions. A significant change has also been made in the area of novelty search, which may affect the procedures required for inventions in chemistry and pharmaceutical sciences. Additionally, facilitation is on the way for demonstrating the level of inventiveness, in particular for biotechnological and pharmaceutical inventions.

The number of changes is enormous and exceeds the scope of this post. Below is a short presentation of selected changes.

[...]

At present, the European Patent Office is working on clarifying the issue of patentability of inventions which were previously considered non-technical. This is relevant because recognizing the technical character is the first and foremost condition for an invention to be considered as such. The examination of the novelty and inventive step cannot begin until the first condition has been met.

Inventions that were previously denied technological character were the so-called computer implemented inventions (CIIs). Last year the EPO decided that those inventions are in fact technical. Artificial intelligence-based inventions (AI) and machine learning-based inventions (ML) have recently followed suit.

In the previous versions of its guidelines, the EPO demonstrated that it should be assumed that AI/ML inventions are non-technical. In the new version of its guidelines, the EPO has changed that approach and stated that the EPO's experts must assume that AI/ML-based inventions may have technical character.

Although this change may seem only superficial, it offers a significant improvement for the applicants. To be more precise, after the guidelines come into force, what the EPO experts will have to demonstrate is a lack of technical character of an AI/ML-based invention, whereas earlier it was the applicant who had to prove that the invention had technical character.

It is worth noting that there is a chance that the procedure for AI/ML inventions will become even friendlier for applicants this year. The European Patent Office is currently deliberating whether inventions based on computer simulations can be patented. Should this happen, it will become easier for applicants to not only demonstrate the technical character but also to demonstrate that this type of invention involves an inventive step.


So just like that they ignore caselaw, violate the EPC, throw away instructions from Parliament and trample over software developers (who were never consulted about this).

Lobbying by the litigation industry, helped by their media (with buzzwords and hype waves), may have yielded results.

"Lobbying by the litigation industry, helped by their media (with buzzwords and hype waves), may have yielded results."The EPO is basically stepping away further and further -- even more so under Campinos -- from the rules that govern it. Then it goes to other countries for photo ops that yield this kind of piece from J A Kemp LLP (patent litigation firm). To quote: "It has been announced that the European Patent Office (EPO) has signed an agreement with the government of Georgia to enable European patents to be validated in Georgia. The validation agreement will enter into force once it has been adopted into Georgian law."

So EPO guidelines become another country's too? Even outside Europe? Even if the EPO violates the law? We don't suppose Georgia's 'IP' people understand that nowadays many European Patents are fake/invalid patents. We've also just noticed this new puff piece from World Intellectual Property Review, reminding us once again that it's little but a megaphone of litigation zealots and -- by extension -- EPO management. To quote: "The European Patent Office (EPO) has released new search tools in a bid to improve the world’s largest free collection of patent documents. In an announcement yesterday, November 19, the EPO said the new Espacenet is a “substantially revised and improved version” of its existing patent information search tool. It said new functions will make it easier for users to conduct searches and access more than 110 million patent documents from across the world for free."

"Good luck to the lone inventor, searching monopolies or millions of submarine patent ambushing him/her, awaiting litigation opportunities to bankrupt his/her business."For free!

Good luck to the lone inventor, searching monopolies or millions of submarine patent ambushing him/her, awaiting litigation opportunities to bankrupt his/her business. Good for innovation?

Why aren't there BILLIONS of documents? Not yet? Maybe trillions? That would be lots and lots of "innovation"... correct?

We've meanwhile also learned from this self-promotional article in Lexology that "a recent European Patent Office (EPO) case has shown that further measures may well be needed" to "keep information truly confidential..."

But wait, isn't patenting all about publication?

Terence Broderick from UDL Intellectual Property [sic] writes about T2239/15, which concerns MPEG. These MPEG patents are very likely bogus software patents (geometry, mathematics etc.) that are grouped in massive numbers to make it far too expensive to invalidate them all. The EPO should not grant any of these, but in practice it even offers special awards to the person who's responsible for hundreds of these. From the article:

In recent case T2239/15 the EPO considered whether documents which were said to be ‘private’ were also ‘confidential’, in the absence of any agreement to say so.

Two prior art documents were cited in examination proceedings which the applicants claimed were confidential working documents, circulated as part of the MPEG (Motion Picture Expert Group) working group. The applicants submitted that a confidentiality agreement was in place within the working group.

A variety of submissions were filed to support this stance which centred on the secrecy associated with national standards bodies, obligations set out in guidelines for delegates and files which set out that documents resulting from meetings of the working group (known as input documents) were considered ‘private’.

However, the flexible nature of the working group couldn’t support an obligation of confidentiality on the members. Indeed, the documents themselves even indicated that members of the group were encouraged to seek external expertise. It was said that the number of members was indefinite and that no absolute obligation of confidentiality existed.

Therefore, the problem wasn’t sharing documents, but that it couldn’t be guaranteed that all members of that group were covered by an explicit obligation of confidentiality — even if documents shared between the members are accepted as being ‘private’.


The EPO does not care about confidentiality; the EPO violates confidentiality and then covers it up. It's a crime, sure, but if the EPO hides evidence of it, will it count? And even if it got caught, nobody would be punished because of immunity.

Recent Techrights' Posts

Linux and the Freedom Paradox
Linux is losing freedom if some external actors who only use Microsoft tools for development wrest control
Watch the FSF Party Live (via Livestream)
It's in WebM format, which is widely supported by now
Advocacy of Software Freedom Changed, LUGs Became Less Relevant
The way we see it, support groups like LUGs sort of outlived their usefulness when it became easier to install GNU/Linux
For the Second Time in a Few Weeks Microsoft Lunduke Makes False Accusations Against Senior Red Hat Staff to Incite a Despicable 'Troll Army'
Nothing that Microsoft Lunduke claims or says can be trusted
Compromised by NVIDIA Proprietary Library
Meanwhile in Boston there are "[r]oundtable talk with FSF volunteers (both in-person and online)"
How Software Patents Were Viewed or Their General Status Changed Over Time
A rough summary
 
Links 05/10/2025: Lufthansa Layoffs (4,000) and More Spotify Woes (Aside From Massive Debt)
Links for the day
The Free Software Foundation's Livestream Has Ended, Video/s Might be Online Next
I've asked whether they'll upload video of some of the event; I still wait for an answer
The Register MS Does Not Know the Difference Between Microsoft GitHub and GitLab
At the time of writing (October 5) the article from "Thu 2 Oct 2025" remains uncorrected
"Bullshit Generators" (What RMS Calls LLMs) and Fake Images Already Target the FSF
Why does Google News promote fake articles about the FSF while omitting all the real ones?
Software Patents as a Bubble
Don't invest resources in hype; if you detect a bubble, run away from it
Links 05/10/2025: Political Leftovers, Climate Change, and Security Incidents
Links for the day
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Saturday, October 04, 2025
IRC logs for Saturday, October 04, 2025
When Microsoft "Integrates" Something With "AI" It Means It's Losing Money and Is Generally Hopeless
how did Bing fare after 36 months of LLM slop being hyped up as "replacement" for search?
Most Certificates Don't Improve Security, They Mostly Increase Downtime (for No Good Reason)
The 'Gemini sites' (capsules) are a growing force
The statCounter Site Has Data Integrity Problems
Maybe we'll get back to statCounter when its data becomes more "stable" again
10 Ways to Combat Software Patents
software patents are loathed also by proprietary software developers
"Just a Little Bit of Meat..."
Free software "absolutism" is not a radical stance, more so if the only "radical" belief the user possesses is that he or she must be in control of his or her software, and by extension his or her computer
Red Hat is Ignoring the Free Software Community, It's a "Fortune 1000" Vendor
Red Hat's blog also participates a lot in promoting of Wall Street's latest pump-and-dump "AI" scheme
Free Software Foundation Party Has Begun
We shall be focusing a lot on software patents today
Former Head of the Federal Trade Commission (FTC) Lina Khan Knows Whatever Microsoft Touches Will Die
Just like Skype (as recently as months ago) [...] When Microsoft grabs things, or when it buys things, it almost never ends well
Slopwatch: Fake Articles About LibreOffice in Austria and Wine 10.16
very short
Links 04/10/2025: "attempted Coup" Noted in Facebook, Russia Kills Journalists via Drones
Links for the day
Gemini Links 04/10/2025: Anesthesia and Baudpunk
Links for the day
Links 04/10/2025: "Privacy Harm Is Harm", Criticism Outlawed in US
Links for the day
Garmin Uses Linux for Some of the Garmin Products, Now It's Sued by Strava Using Software Patents
Software patents should never have been granted in the first place
Richard Stallman Will Give a Talk in Sweden in 6 Days
Dr. Stallman, despite his battle with cancer is still alive and mentally sharp
FSF Turns 40
We'll be focusing on patent-related topics this weekend
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Friday, October 03, 2025
IRC logs for Friday, October 03, 2025
Gemini Links 04/10/2025: Distro Hopping and "Part Time"
Links for the day
We Are Turning 19 in One Month, FSF Turns 40 in 3 Hours (CET)
For our anniversary next month we still have no concrete plans
Patent Docs (or PatentDocs) Learned the Wrong Lessons From the Death of TypePad
Had they gone ahead with an SSG, they'd become a lot more future-proof
USPTO Patent Bubble Already Imploding, After Decades of Artificial Inflation, Entire Offices Close for Good
we can deduce that financial pressures (lack of "demand" for monopolies) play a role
TikTok is Not Harmless (Being CheeTok in the US Will Advance Orange Agenda)
Social control media isn't "fun and games"; it's a digital weapon that lets hostile groups or nations infiltrate others, then turn them against themselves
Andy Farnell and Helen Plews Explain What "Modern" Tech Does to Old People
Imposing terrible tech "religion" on people is not helping them
Tomorrow the Free Software Foundation (FSF) Turns 40 and Its Web Site is Still Slow Due to DDoS by LLM Slop Bots
For an advocacy group, uptime is important (for its message to remain accessible)
Slopwatch: Google News as a Firehose of LLM Slop About "Linux"
Google News is really bad
Datamation, Where I Used to Publish Articles, Appears to Have Been Sold to TechnologyAdvice Only to Become a Slopfarm
I'd prefer to not associate with that site anymore
Links 03/10/2025: "NPR’s Economics Lessons Come With Neoliberal Spin" and Canada Post at Risk
Links for the day
Gemini Links 03/10/2025: Panic Attacks and Food Adulteration
Links for the day
Links 03/10/2025: Lawyers Caught Using LLM Slop Explain Why They Did It, LibreSSL 4.1.1 and 4.0.1 Released
Links for the day
FSF Board Grew 50% Since Last Year, Has New President, Turns 40 in Two Days
It's a good move for the FSF and - by extension - for software freedom
Links 03/10/2025: Conflicts, Death of TypePad, and TikTok/CheeTok Gives a Boost to Far Right Groups in Europe
Links for the day
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Thursday, October 02, 2025
IRC logs for Thursday, October 02, 2025
Slopwatch: Linux Journal, Google News, and LinuxSecurity
They carry on polluting the Web with fake articles
Gemini Links 02/10/2025: Kubernetes With FreeBSD and robots.txt
Links for the day