Matt Bodie

The Yale Law Journal Forum recently hosted a collection of essays under the rubric of “Reimagining and Empowering the Contemporary Workforce.” Two of these works deal specifically with the challenges for workers posed by the dramatically reoriented data landscape: Data Laws at Work by Veena Dubal, and AI and Captured Capital by Ifeoma Ajunwa. Both essays are essential reading for those interested in data protection and regulation within the workplace.

Professors Dubal and Ajunwa present a nice contrast in their approaches to empowering workers within the data revolution. Dubal follows a more traditional privacy-oriented approach, seeking to further restrict employer access to, and use of, employee data through narrow permissions and harsher penalties. Ajunwa, on the other hand, argues that worker data represents business capital, and she contends that employees should have long-term rights to the value generated from that data. These two avenues—inalienability restrictions and property rights—should both play bigger roles in our system of workplace data regulation, especially within the world of algorithms and artificial intelligence.

For American audiences, the European Union’s approach to data regulation may seem like an unattainable dream of privacy protection. Since 2018 the General Data Protection Regulation (GDPR) has required justification for the processing of someone’s data, and has offered a panoply of rights to data subjects, such as notice, portability, rectification, and erasure. Dubal’s theme in Data Laws at Work, however, is that even this idealized set of protections ultimately fall short in the context of the workplace.

For those looking to understand the current state of European protections, Dubal provides thoughtful reviews as to the employment data ramifications for the GDPR as well as two EU initiatives from the past year: the Artificial Intelligence Act and the Platform Work Directive. As Dubal describes, the AI Act requires that workers be directly informed about the use of AI systems and bans the production and use of AI systems that emotionally manipulate people, while he Platform Work Directive provides a set of specific rights to these workers, including and transparency obligations, the prohibition of processing certain types of data, and required impact assessments for automated decision-making systems.

These employment data regulations are clearly more robust than those in the rest of the world, including the United States. But Dubal does not believe the EU’s approach is sufficient to protect workers from the depth and breadth of Big Tech’s reach. In the clutches of new systems of massive data collection and analysis, workers lack the independence to exercise individual data rights meaningfully. Static models of scientific management have given way to dynamic models of algorithmic management that assess workers on a collective and relational basis. As a result, rights of notification or rectification do not properly empower workers, as the information is difficult to understand and must be challenged collectively.

Instead of individual causes of action, Dubal argues for stronger affirmative requirements: disclosure obligations, periodic impact assessments, and independent third-party data audits. She also makes the case that policymakers should impose outright bans, rather than just disclosure or limitations, on employer tools such as algorithmic decision-making as to wages, discipline, and termination. As Dubal concludes, “data laws focused on the workplace must affirmatively proscribe—not merely elucidate—these forms of worker control.” (Dubal, p. 447.)

In AI and Captured Capital, Ajunwa similarly sees worker vulnerability to machines of mass analysis and the companies that control them. At present, workers are feeding these particular beasts through “captured capital”—her description for the coercive collection and use of worker data that further refines and expands the algorithms that may eventually replace them. Worker data, Ajunwa argues, should be considered capital, as it is used to create value and must be considered valuable on its own, and that capital should be owned by workers, rather than the entities that employ them.

As Ajunwa describes, workers are uniquely disempowered with respect to their data under current law: default rules assign property rights to the firms; automated systems are developing at a surprising rate and may soon overtake people-based productivity in many fields; and labor rights are unenforceable in the face of boundary-less global labor markets. Her picture, like Dubal’s, is fairly grim.

Ajunwa, however, takes a different approach to fighting back. Rather than beefing up privacy protections, she advocates for the creation of ongoing property rights for workers in the value that their data generates. She proposes three different potential models: (1) worker data as “stake capital,” providing governance rights to workers similar to angel investors or venture capitalists; (2) a communal data-licensing regime similar to agreements between actors and film studios for the use of their likeness, image, or voice; and (3) an ongoing source of stable funding, similar to a universal basic income, to be administered by the International Labor Organization or other NGO. Logistical and organizational hurdles confront all of these proposals, but each takes a real-world example and applies it to ameliorate the economic and power imbalances of the data-centric economy. These efforts to redress the harms from captured capital would “ensure that workers regain some measure of control over their data and can benefit from the data they create for firms.” (Ajunwa, p. 404.)

Employee data protections have too long been mired in the paradigm of privacy protections based on liability for individual harms. A worker can sue if the employer breaks open a locker without permission or exigent circumstances, or sets up a secret camera in a private area, or intercepts a personal phone call without notice. But the ongoing campaigns of data incursions and degradation, cogently described by Dubal and Ajunwa, are largely allowed even in the most restrictive of jurisdictions.

Both authors present novel and impactful ideas for reforms: Dubal supporting prohibitions rendering certain personal data inalienable, and Ajunwa providing property interests to workers in the ongoing processing of their information. These proposals have the potential to change the data dynamics within employment relationships in meaningful ways. Kudos to these talented scholars for their insightful and important contributions on data in the employment relationship.