BASED on the world's trashiest tech tabloid, ZDNet [1], there are patents based on "HEY HI" (AI). The USPTO uses that as an excuse for granting abstract patents, with WIPO (UN) cheering from afar. Then we have TechDirt [2], and Slashdot linking to TechDirt [3], which wrongly described this as an "EU" thing yesterday (to be fair to TechDirt, it did cover some EPO scandals in the Battistelli era, it opposes software patents in Europe, supports 35 U.S.C. ۤ 101 and so on). It spoke of the "EU Patent Office" and more repetition of the falsehood (EPO predates the EU and has different member states). Those three examples from yesterday are good examples of bad and misguided journalism in this domain (truer than ever before so far in 2020).
"Unless we can get more people to accurately assess and fact-check, we'll carry on drowning in a sea of lies and myths."António Campinos has fallen in love with "AI" as a buzzword; but he apparently wasn't happy to allow computer-generated applications (or the likes of that, gaming the system with automation at the applicant's end).
Sadly, debates about these issues are rarely honest. This past week I responded to many misleading or outright false tweets from the EPO. Tweets aren't generally known for high quality, but those ones were just utter lies, sometimes paid-for lies.
A couple of weeks ago Hartwig Thomas (Switzerland) entertained one such EPO tweet when he responded with: "The claims are all false/heavily slanted:
- Disclosing some technique is almost never “unsafe”!
- The concept of “owner” is related to the misused concept of “property”.
- Open communication of a technique or method or machine does NOT prevent its commercial exploitation (just ask Red Hat!)
- Why should use of knowledge ever be “authorized” by anybody?
- Read boldrin/levine “against intellectual monopoly” who illustrate that patent law fails to achieve its goals and hurts society."
That's just a response to one EPO tweet among many (there are hundreds of new ones each month). Unless we can get more people to accurately assess and fact-check, we'll carry on drowning in a sea of lies and myths. ⬆
Related/contextual items from the news:
We've written a bunch about why AI generated artwork should not (and need not) have any copyright at all. The law says that copyright only applies to human creators. But what about patents? There has been a big debate about this in the patent space over the last year, mainly lead by AI developers who want to be able to secure patents on AI generated ideas. The patent offices in the EU and the US have been exploring the issue, and asking for feedback, while they plot out a strategy, but some AI folks decided to force the matter sooner. Over the summer they announced that they had filed for two patents in the EU for inventions that they claim were "invented" by an AI named DABUS without the assistance of a human inventor.
[...]
The problem, yet again, is in people's misguided belief that everything must be owned by someone, and that somehow without a patent it is impossible to successfully commercialize or market a product. There is tremendous evidence to the contrary (including just by looking at products after their patents run out -- which is often a time when more innovation occurs, since there's greater competition driving improvements). But, instead, you hear nonsense like the following from Prof. Ryan Abbott, who helped file the two now rejected patents, arguing that without patents, somehow these inventions might not come to be: