It is now understood that machine learning algorithms can produce unintentionally biased results. For the last few years, legal scholars have been debating whether the disparate treatment or disparate impact theories available under Title VII of the Civil Rights Act are capable of protecting against algorithmic discrimination. But machine learning scholars are not waiting for the legal answer. Instead, they have been working to develop a wide variety of technological “fairness” solutions that can be used to constrain machine learning algorithms. They have discovered that simply blinding algorithms to protected characteristics like sex or race is insufficient to prevent algorithmic discrimination. Given enough data, algorithms will identify and leverage on proxies for the protected characteristics. Recognizing this, some scholars have proposed “fairness through awareness” or “algorithmic affirmative action” — actively using sensitive variables like race or sex to counteract unidentified sources of bias and achieve some mathematical measure of fairness in algorithmic decisions. But is algorithmic affirmative action legal? This article is the first to comprehensively consider that question under both Title VII and the Equal Protection clause of the Fourteenth Amendment. The article evaluates the legality of the leading fairness techniques advanced in the machine learning literature, including group fairness, individual fairness, and counterfactual fairness. The article concludes that existing affirmative action doctrine under Title VII and existing constitutional equal protection jurisprudence leave sufficient room for at least some forms of algorithmic affirmative action.

When is paternalistic regulation of risky work justified, and who should get to decide that question? The prevailing economic account of OSHA regulation is that market interventions are justified only by spillover and informational market failures. This article challenges that prevailing account by contending that worker safety regulations are also justified where necessary to enforce the relational expectations of employees, which are shaped not only by formal contract but also by social norms and norms internal to the employment relationship. Drawing on the relational theory of contract, this article identifies an underappreciated purpose for employment law: anti-opportunism. If preventing relational opportunism is a valid purpose for workplace safety law, then OSHA regulation is justified in a broader range of circumstances than has been generally assumed under more traditional analyses. This article contends that regulation is justified where necessary to prevent employers from engaging in self-interested behavior that runs contrary to the parties’ relational expectations. Applying this standard is a highly context-specific, fact-dependent task, which makes Congress' decision to delegate enforcement authority to OSHA appropriate.

Three possible approaches have been advanced for setting appropriate occupational health standards: the feasibility principle currently employed by OSHA, cost-benefit analysis, and a “soft” cost-benefit analysis that allows for qualitative considerations. This Article rejects all three and advances a fourth possible approach that would focus on counteracting “health theft” by employers — employer actions that expose workers to health risks without compensation. An anti-health theft approach recognizes that Congress’s purpose in enacting the Occupational Safety and Health Act was to maximize worker welfare, not to maximize overall total social welfare. This Article makes both the positive claim that counteracting health theft was Congress’s intent when it enacted the OSH Act and the normative claim that an anti-health theft approach sets a rational regulatory standard justified on distributional grounds. The Article urges OSHA to adopt the anti-health theft principle as a reasonable interpretation of the ambiguous Section 6(b)(5) of the OSH Act.