R Street Tech, January 2020
I GOT 230 PROBLEMS BUT USMCA AIN’T ONE — In the Daily Caller, Jeff talks about the recently signed U.S.–Mexico–Canada trade agreement, and in particular a provision in it that contains protections for online platforms. The USMCA provision is modeled after Section 230 of the Communications Decency Act, which has come under scrutiny recently, but Jeff argues that the provision is “a tremendous win for users” by protecting Internet companies from “foreign government jawboning” and keeping content moderation “based on market and consumer interests rather than political pressure from Washington.”
DO NOT USE IF ALLERGIC TO FREE TRADE — Bill Watson and Clark Packard published a white paper on biologics provisions in U.S. trade policy. After noting the competitive benefits of a robust market of generic drugs and biosimilars, the paper considers whether certain exclusivity periods that delay these competitive markets are appropriate subject matter for free trade agreements and concludes that “long biologic exclusivity provisions fit perfectly into the category of misguided intellectual property provisions” that “would not open markets or even promote innovation.”
THE ROBOT ATE MY HOMEWORK — In comments filed with the U.S. Patent and Trademark Office, Caleb Watney explains why copyright law needs to accommodate new technologies such as artificial intelligence. Training AI systems requires vast amounts of data, much of which is potentially subject to copyright, meaning that AI companies often must rely on exceptions to copyright law, in particular one known as “fair use.” To the extent that there is uncertainty in the fair use doctrine, then, costs of litigation become a concern for startups; “the current ambiguity surrounding the fair use exemption disproportionally hurts smaller firms.”
ALL YOUR COMPUTER SCIENTIST ARE BELONG TO US — Caleb appeared on Libertarianism.org’s Building Tomorrow podcast, where he discussed the role of immigration policy in new technology development. Because of the complex, lotteried nature of immigration visas for highly skilled laborers, experts in fields such as artificial intelligence are pushed away from startups; students are “almost twice as likely or more likely to end up working at a big incumbent firm.” Caleb was also on the Federalist Society’s Regulatory Transparency Project podcast, talking with Adam Thierer about driverless cars.
I GUARANTEE HE IS 100% REAL — Jeff Westling submitted a statement to a House committee hearing on “Manipulation and Deception in the Digital Age.” In it, he explains how deep fake videos—the realistic AI-generated videos that gave rise to the hearing—need to be considered within the historical context of “a long line of tools and techniques used for deception.” Incautious responses to deep fakes, he explains, can “limit the numerous beneficial uses the technology can provide” and “shift attention away from more rudimentary forms of disinformation that better exploit the psychological factors driving trust and sharing online.”
WARNING: IMPLICIT CONTENT — In the Morning Consult, Daisy Soderberg-Rivkin writes that content moderation does face a bias problem, but not the one everyone is talking about. Rather than anti-conservative bias, which has been the subject of multiple hearings in Congress, Daisy explains that “implicit bias,” a well-studied psychological phenomenon, has important consequences for content moderation. There are steps that companies and policymakers can take to reduce these problems, she explains, but they first require a “radical shift in the political conversation relating to bias in content moderation from mythical issues to ones that affect a large portion of our society.”
WHO’S COPYING WHO NOW? — For the last decade, Oracle has argued in litigation that Google committed illegal copyright infringement by copying a part of Oracle’s software called an API. In an article published in Ars Technica, Charles Duan argues that if Oracle is right, then it must be a copyright infringer too, because Oracle copied an API from Amazon. The better answer though, Charles explains, is that no one is infringing any copyrights, because API copying is common across industries and, more importantly, is necessary to a competitive technology market.
COPYRIGHT ALL THE THINGS — Expanding on that argument, Charles authored an amicus curiae brief to the Supreme Court, in conjunction with Public Knowledge and the Niskanen Center, explaining how API copying is so prevalent that it arguably occurred with flags at the Battle of Trafalgar, century-old army telegraph codes, hospital emergency pages, federal drone regulations, the World Meteorological Organization and much more. These numerous examples show, he concludes, that Oracle’s copyright theory cannot be correct, or else every company and even the federal government would be a copyright infringer.
SOUND OFF, ONE TWO — Should a patent on audio speaker systems require the federal government to block imports of laptops, tablets and smartphones? Bill Watson and Charles Duan think not, and make that clear in a comment filed with the International Trade Commission in a dispute between Sonos and Google. They argue that Sonos’ patent has nothing to do with these general-purpose devices that have many uses beyond speaker systems. To exclude them from importation “would harm competitive conditions and consumer welfare without furthering any interest in the enforcement of Sonos’ patent rights.”
JUMP ON THE (BROAD)BAND-WAGON — At the Opportunities for Bipartisan Tech Policy 2020 conference hosted by Next Century Cities, Tom Struble spoke on a panel about “Broadband Access vs. Broadband Subscriptions: The Difference and Why It Matters.” Tom explained how access and adoption are two sides of the same coin, and that policies supporting one side of the market also help support the other. For ways to promote adoption among users who have access to broadband but choose not to subscribe, Tom suggested that targeted regulatory reforms can help drive down subscription costs while community-learning programs can help promote digital literacy and broadband use among elderly and disadvantaged populations who could benefit tremendously from broadband access if they only knew how to use it.