Law and Political Economy: Labor, Social Control, and Counterpower

This conference will take place from March 31 to April 2, 2023, at Harvard Law School in Cambridge, Massachusetts. It will feature several panels on technology. If interested, please contact Elettra Bietti for more information. 


NYC Mayor Eric Adams put forth a guidance calling members of the public to “lower their face masks to reassure store workers they’re not criminals.” This policy elicited a considerable stir among the community, as it implicated not only public health concerns but also issues of privacy—given Adams’ rationale that removing face masks would afford a better chance for security cameras and law enforcement to “identify criminals.” Against the backdrop of this contention, this policy choice raises serious questions about surveillance, identification, and their discriminatory impact on communities of color. 

The FTC is scaling up its efforts to investigate Twitter’s data and privacy practices, focusing on “whether Twitter has adequate resources to protect its users’ privacy after mass layoffs and budgets cuts” ordered by Elon Musk. As a part of its investigation, the commission is seeking an interview with Musk. FTC is one of a handful enforcement agencies that began scrutinizing Twitter after Musk took over leadership. In response to criticism that the FTC was launching an aggressive campaign to harass Twitter, FTC spokesman Farrar countered that “protecting consumers’ privacy is exactly what the FTC is supposed to do.” 

Gigi Sohn, President Biden’s nominee to the FCC, has withdrawn her nomination. Sohn was re-nominated as the third Democratic commissioner at the start of this Congress, after a party-line split on her nomination last year. Her withdrawal comes after a string of opposition to her nomination to the post, including Senator Manchin who announced that he would vote “no” over “years of partisan activism” and Sohn’s alleged alignment with the movement to defund the police and limit police surveillance tools. 

The FTC has proposed banning BetterHelp, Inc., an online counseling service, from “sharing consumers’ health data, including sensitive information about mental health challenges” for targeted advertising. This proposed order also requires BetterHelp to pay $7.8 million for charges revealing that the company shared consumers’ sensitive mental health data with third parties such as Facebook and Snapchat, despite promises of data confidentiality. For further discussion, see the following paper by Joanne Kim, also linked under “Papers” below: Data Brokers and the Sale of Americans’ Mental Health Data. 

Digital payment systems have taken hold of about 99% of the adult population in India, who have adopted a biometric ID number. Deep-rooted reliance on this digital infrastructure intensified in light of the pandemic, as the Indian government “used ID numbers to manage the world’s largest vaccination drive and deliver financial aid.” This trend proves a salient behavioral shift in what was primarily a cash-driven economy.  

The education ministry of Poland has announced its plan to mandate annual physical fitness tests in schools for children in ages 10 and upwards. This proposal will be followed by the aggregation of resultant data in a national database called Sportowe Talenty (Sporting Talents), administered by the sports ministry. In addition to the Institute of Sport, this database will be analyzed and results shared with relevant ministries to inform policymaking in Poland. 


Data Brokers and the Sale of Americans’ Mental Health Data by Joanne Kim – This paper discusses in depth the sale of sensitive mental health data in the context of the data broker industry and calls for clearer policies for consumer privacy protections in the U.S. 

“Provable Copyright Protection for Generative Models” by Boaz Barak – This post introduces a paper providing “a formalism that enables rigorous guarantees” on the similarity and lack thereof between “the output of a generative model and any potentially copyright data in its training set.” The research uses both language (transformer) and image (diffusion) models to build algorithms that can provide a training pipeline with minimal degradation in efficiency and quality of output. If interested in further discussion, please reach out to

(Compiled by Student Fellow Stephanie Shim)