11.11.18
Gemini version available ♊︎Automation of Searches Will Not Solve the Legitimacy Problem Caused by Patents Lust
Summary: The false belief that better searches and so-called ‘AI’ can miraculously assess patents will simply drive/motivate bad decisions and already steers bad management towards patent maximalism (presumption of examination/validation where none actually exists)
THE emergence of SCOTUS‘s decision on Alice and today’s 35 U.S.C. § 101 was quite revolutionary. We have no issue with USPTO-granted patents on physical things (an example from several hours ago can be seen here) but on algorithms — something which the European Patent Office (EPO) too has been guilty of lately.
“Examiners need to better understand and respect patent scope, irrespective of what was done in the past (prior art).”Recently there were those who framed prior art [1, 2] — not patent scope — as a core issue. Not even Watchtroll was happy about it (labeling it “An Overstated Solution to Patent Examination”). Examiners need to better understand and respect patent scope, irrespective of what was done in the past (prior art). Compare this to prior Watchtroll rants/coverage about prior art (like this from 2 days earlier).
We don’t mean to say that prior art never matters; alluding to this failing company (GoPro), for instance, Patently-O recently highlighted this case where prior art eliminated wrongly- and already-granted US patents, thanks to PTAB’s availability/existence. To quote:
On appeal, the Federal Circuit disagreed with the PTAB’s conclusions — holding that the Board too narrowly “focused on only one of several factors that are relevant to determining public accessibility in the context of materials distributed at conferences or meetings. . . . [O]ur case law directs us to also consider the nature of the conference or meeting; whether there are restrictions on public disclosure of the information; expectations of confidentiality; and expectations of sharing the information.”
After reviewing the matter Federal Circuit rejected the PTAB analysis and found that the catalog’s use at the show counted as prior art.
“My win in the GoPro v. Contour IP case is the subject of Dennis Crouch’s Patently-O post,” Professor Lemley wrote in Twitter (about the above). It is worth noting that Watchtroll is proudly promoting Iancu’s war on PTAB. Like Battistelli, Iancu hopes that the patent legitimacy problem will go away by marginalising the auditor.
“Search has always involved a degree of so-called ‘AI’ without it being exploited as a buzzword; this is becoming ridiculous.”Speaking of prior art, there’s no way to automate search for it as well as assessment. Domain experts are needed, hence the role of examiners. Days ago we saw the report “North Side company secures $250K to help develop AI-powered patent search tool”; They said “AI”, so someone gave them money for something that would work poorly (if at all). IPPro Patents coverage of Dennemeyer buying Octimine also used the term “AI” in the headline. Search has always involved a degree of so-called ‘AI’ without it being exploited as a buzzword; this is becoming ridiculous. To quote:
The Dennemeyer Group has acquired patent search service provider Octimine Technologies, a start-up based in Munich, Germany.
Octimine, which was founded in 2015 founded by former LMU Munich and Max Planck Institute scientists Michael Natterer, Matthias Pötzl and Dietmar Harhoff, provides patent-searching and analysis tools utilising artificial intelligence (AI).
Judging by the everyday nonsense we see as early as this morning, there’s “AI” nonsense everywhere. The latest fashion/trend? Marketing? António Campinos uses it all the time to promote software patents in Europe. █