Do Androids Dream of the Common Good? Legal Scholars Probe AI’s Public Interest Failings

AI Through a Legal Lens: Beyond Efficiency to the Common Good

A growing body of legal scholarship is reframing the conversation around artificial intelligence, shifting the central question from technical performance to a much older one: does this technology serve the public interest? The inquiry, crystallized in a recent digest of academic work, asks whether existing laws can hold AI systems accountable when they shape consequential decisions and whether the current regulatory architecture is enough to keep deployment oriented toward the collective welfare rather than just corporate efficiency or innovation.

The analysis, framed explicitly as legal scholarship rather than a policy white paper or industry report, examines how courts, lawmakers, and administrative agencies interpret the social impacts of AI. Instead of focusing on benchmarks or compute scaling, the work evaluates the normative frameworks through which AI’s risks—discrimination, opacity, due process violations, and privacy erosion—are filtered into legal doctrine. At its core lies the concept of the “common good,” a principle that demands AI governance prioritize broad societal benefit over narrow economic gain.

Accountability and the Limits of Current Law

One of the central themes is the question of responsibility when AI systems cause harm. Scholars are dissecting whether existing liability regimes—tort law, product liability, anti-discrimination statutes—are adequate for a technology that can act unpredictably, make errors invisible to human review, or replicate bias at scale. The digest highlights a fault line in legal academia: some argue that well-established doctrines can be stretched to cover algorithmic injuries, while others insist that the vacuum of clear rules creates accountability gaps that only new legislation can fill.

This tension is especially visible in sectors like criminal justice, hiring, and credit scoring, where automated decisions can deprive individuals of liberty or economic opportunity. Legal critics warn that without binding public-interest constraints, developers and deployers may treat fairness, transparency, and due process as optional features rather than foundational requirements.

Global Governance Experiments Offer Reference Points

The scholarship contextualizes these debates within a patchwork of emerging regulatory frameworks. The European Union’s AI Act, which classifies applications by risk and imposes strict obligations on high-risk systems, is frequently cited as an attempt to hardwire the common good into market rules. Meanwhile, the United States has leaned on guidance from the National Institute of Standards and Technology, whose AI Risk Management Framework outlines voluntary processes for mapping, measuring, and governing AI risks. On the international stage, the OECD AI Policy Observatory tracks how different jurisdictions define principles like transparency and accountability, offering comparative data that legal scholars use to argue for or against harmonisation.

These instruments, the scholarship implies, are not just technical blueprints—they are legal artefacts that encode a particular vision of the public good. Examining them through a juridical lens reveals unresolved dilemmas: Can a voluntary framework truly protect rights? Does a risk‑based regulation capture the cumulative impacts of AI on marginalised communities? And who gets to define what “common good” means when the technology is deployed across borders with conflicting legal traditions?

“The common good is not a design specification that engineers can simply tick off. It is a legal and political commitment, and our current laws were never written with autonomous systems in mind.”

This perspective, which runs through the digest, underscores a broader call for legal innovation. Some scholars advocate for mandatory algorithmic impact assessments, others for a new federal agency with the expertise to adjudicate AI‑related disputes, and still others for an overhaul of anti‑discrimination law so that proof of disparate impact on protected groups triggers immediate remedial duties, even when the offending system is opaque. The European Commission’s AI policy and AI Act materials offer one model, but the digest makes clear that for many American legal thinkers, the path remains contested.

From Niche Debate to Mainstream Urgency

The Substack‑based digest is a symptom of a larger shift. As generative AI and agentic systems penetrate sectors ranging from healthcare to public benefits administration, the legal academy’s abstract debates over “sociotechnical regulation” are suddenly concrete. Courts are already seeing challenges to algorithmic eviction scoring, automated welfare fraud detection, and AI‑driven teacher evaluations—cases that force judges to interpret old statutes in new contexts. The scholarship reviewed in the digest provides judges and regulators with doctrinal tools to resolve these disputes in a way that aligns with democratic values rather than leaving the outcome to the terms of service of a private company.

For the public, the takeaway is that the “common good” framing is not a philosophical abstraction. It translates directly into questions of whether a loan applicant can see and correct the data used to reject them, whether a criminal defendant can cross‑examine the code that ties them to a crime scene, and whether an employee displaced by an algorithm has any legal recourse at all. Legal scholars are increasingly insisting that these are not just technical glitches—they are failures of governance that law, and only law, can repair.