Thursday, August 11, 2022

IPSC, Stanford: Opening Plenary Session: Under-Represented Groups in the IP System

Representation of Female Inventors on Patent Teams

Jordi Goodman

Equity would be achieved in 2092 if current trends continue. Group dynamics: is everyone equally likely to be on a team? E.g., if women are 20% of STEM workforce, and teams are equally likely, then two person teams should be all female .2x.2 of the time, or 4%. In small groups, though, women are underrepresented; in larger groups of 5 all-male and all-female teams are overrepresented. DEI initiatives may be pushing women together to work and not promoting working together? In university groups, the most underrepresented group is a mixed one (e.g., 2 men and 2 women).

Curating Black Music: Copyright, Ownership & Commodification

Olufunmilayo Arewa

Book project: How recording business has shaped identities through curation, including of sexuality and of profits. Copyright incentives to create v. incentives to exploit. Black musical forms are at the core of pop music, not just in the US but globally. African-American artists are particularly affected by sharecropping model in music where it is difficult if not impossible to get out of debt: widespread manipulation of accounting and lack of transparency, which remain dominant today even after major technical change. Lots of discrimination, including by gender; differential treatment of African-American artists. Racial scripts about creativity are the basis for denigrating artists.

Example: Lil Nas X, Billboard removed Old Town Road from Hot Country chart, claiming it didn’t embrace enough elements of “today’s country music to chart in its current version.” Billboard said it wasn’t race-related, which is not particularly credible; shows the historical curation of country music category as for white people. Other genres are considered white still: folk, rock and roll. “Urban,” R&B, etc. are Black. That limits opportunities and denies historical connections, e.g. country’s roots in Black musical forms. [Banjo/fiddle crossover, history of interracial playing, “Black hillbillies” who couldn’t get record contracts.] This is about curation: we decide Black is a category, what fits in it, and what rights people in that category have. Shaped by conceptions of what culture is or should be, but has an economic impact.

Another example: sculptures. We think of them as white but Greeks and Romans painted them. What does it mean to paint those sculptures? Resistance to doing so reflects imagined histories of whiteness. Current assumptions about race shape our visions of the past, present, and future. They can change—they did in the past, which didn’t think in the same way.

African-American artists traditionally got lower royalty rates and paid out in different ways. Incentives to exploit marginalized groups—record companies can treat them worse.

The Gender Gap in Academic Patenting

Miriam Marcowitz-Bitton, Michael Schuster, and Deborah R. Gerhardt

Mary Cade & June Davis: two women very important in making Gatorade drinkable—saving the project—but not credited in any patent, whereas husband Bob Cade (who vomited when he drank the first version) gets the credit. When Stokely got the rights, it assigned chemist Davis to make it more palatable, and she did, but that didn’t lead to public contemporaneous credit.

In 2015, 29% of patent applications name at least one woman inventor—up from 17% in 1997. 7.2% of named inventors are women. There is also evidence of bias in the examination process against people with identifiably female names. About 63% of academic patents filed by all-men groups; under 2% filed by all-women groups. 64% of inventor teams made up only of women are single-inventor teams; not true of inventor teams made up only of men, where many are 2-person teams. Also seeing citation rates in later patents that favor men.

Computer Software Patents and the Gendered View of Computer Programming as Labor or Innovation

Nina Srejovic

Lots of key women programmers in early computing history. When only women were programming, men thought it was sexier to build hardware and no one thought the software was important. Women were viewed as mere operators of machines built by men. Gov’t rating was SP, for “sub-professional.”

Change in identity of profession: to ensure that computer work would not be “handmaiden,” private companies/ newly established professional organizations/academic disciplines/gov’t funding sought to create programming as a discipline in its own right—software engineering.

Lawyers, scholars and judges both fostered and accepted this view. Now the problems were difficult, intellectually challenging, resulting in innovation/invention. Programming became part of the machine: patentable as “instant hardware.” Creator v. user stereotype: what was unpatentable as labor when women did it became patentable when men did it and were perceived to need incentives/rewards. Challenges the notion that women’s own activities determine the number of patents they’re granted. So patent-related metrics themselves encode biases.

Q: about class: also happens with workers who aren’t credited with things they learn from using the machines.

Q: is the answer more software patents or fewer? Maybe the male model isn’t one to emulate. Credit/recognition doesn’t have to be turned into patents.

A: agree we don’t want more software patents of the sort now being issued. But software patents are better than software copyright.

The Innovation Glass Ceiling: How Women are Penalized for Boundary Spanning Research

Ryan Whalen, Tara Sowrirajan, Sourav Medya, and Brian Uzzi

Key premises: there is a STEM gender gap, including in patenting; atypical boundary-spanning inventions are particularly important because they introduce new product categories and expand the tech/scientific universes by making novel knowledge recombinations. Higher probability of being high-value. [Also high-risk: they are more likely to fail too; but they have higher maintenance percentages when granted suggesting higher value overall.]

But: women face barriers doing this type of research that men don’t. Consistent gap in success when there are multiple CPC subclasses for the claimed invention; where the algorithm can’t determine gender, the claimed invention does as well as it does when the inventor is male. Persists through various explanatory variables like inventor/examiner experience, examiner gender, team size, entity size, etc. Some subclass combinations are typical and others are extremely atypical.

The more atypical the combination and the more of the team that is female, the less the chance of a grant. But the more atypical the combination and the more of the team is male, the better their chances. So men are rewarded for atypicality, and women are penalized.  This also holds for first listed inventor. Similar findings from UKIPO and Canadian IPO.

When women are granted boundary spanning patents, they take over a month longer (with controls), they have fewer independent claims than men, they lose more independent claims—about a quarter—on prosecution than men do.

There is no clarity on causality. Questions: examiner bias, employer bias (e.g., excluding women from teams with great boundary-spanning inventions), patent agent roles, art unit/technical area analysis. Assumption: quality is randomly distributed but we don’t know whether it’s true. [Also I wonder which way that cuts: if women’s low-quality inventions are screened out and men’s aren’t, that could be an even worse problem.]

Q: have you looked for whether men make different atypical combinations than women? A: No.


No comments: