From the EPO Gazette:
Summary: Using bogus interpretations of the EPC -- ones that courts have repeatedly rejected -- the EPO continues to grant bogus/fake/bunk patents on abstract ideas, then justifies that practice (when the audience comes from the litigation 'industry')
YESTERDAY we covered day one of EPOPIC, taking note of overt promotion of software patents in Europe by the António Campinos-led European Patent Office (EPO).
Day two of this event had yet more of the same, namely promotion of software patents in Europe. As people who are programmers probably know already,
"AI" just means software; it's a particular
kind of software, but it's still software. I myself have written algorithms in this domain under more than one employer and more than one university. Despite the fact that we're talking about algorithms here, the EPO is blatantly working around the law again. Yesterday it
wrote: "AI and the patent system: The fundamental concepts of patent law will adapt – from the protection of hardware innovation, to the protection of software innovation."
When they say "patent law will adapt" they mean to say "will be worked around" (or violated) and by "software innovation" they mean programming. The EPO
mentioned "AI" in another context (management of patents themselves), e.g. "Patent information professionals will be able to leverage AI's ability to quickly process massive amounts of data to surface important information. Samuel Davis shares great insights on this topic at #EPOPIC [
...] What are the areas where the EPO uses AI? Automatic pre-classification of applications Automatic CPC classification of patent documents Automatic searches on patent applications Automatic annotation of patent literature Patent Translate..."
It's important to stress that this isn't about software patents but about software being used to
manage patents (we suppose that the EPO will deem such software patent-eligible too).
Methods of violating the EPC (and then ignoring courts that say EPO is violating the EPC) are then
specified by the EPO: "The EPO has developed a methodology in order to determine if computer implemented inventions are patentable: (1) Technical character, (2) Inventive step based on technical features only."
This is nonsense; the nonsensical interpretation above has already been rejected by courts. But
on they go: "EPO expert Argyrios Bailas presents how the EPO is incorporating AI tools in its work and how it examines applications related to artificial intelligence." (combining the two aspects above).
"Patent offices need to provide a predictable, stable & timely approach to the patenting of computer implemented inventions," the EPO
then added.
Examiners should say "NO!" immediately (to such patent applications), as per the law -- a concept that the corrupt EPO fails to grasp. This would be "predictable, stable & timely," to borrow the EPO's own words.
We suppose that the remainder of EPOPIC will have more of these same lies. We shall watch these closely today.
⬆