Cary Coglianese (U Pennsylvania Carey Law) and Oren Perez (Bar-Ilan U Law) have posted “Fighting Risk with Risk” (University of Illinois Law Review, forthcoming) on SSRN. Here is the abstract:

A growing scholarly and policy debate pits advocates of the precautionary principle, who prioritize risk minimization, against proponents of an innovation principle, who argue that regulation suppresses technological progress. Despite their differences, both camps share the assumption that contemporary regulatory law is shaped by a risk-averse orientation that is fundamentally at odds with the risk-seeking ethos of technological entrepreneurship in fields such as artificial intelligence. What this debate overlooks is an alternative approach altogether—one that deliberately embraces risk as a vehicle for risk reduction. Instead of fighting risk by restricting, delaying, or controlling risky activity, the government sometimes justifiably fights risk by embracing risk under legally structured conditions. 

In this article, we introduce and offer a theory of this overlooked strategy. Drawing on examples as varied as so-called ethical hacking used to expose cybersecurity vulnerabilities, prescribed forest burning to prevent catastrophic wildfires, and geoengineering to combat climate change, we explain how fighting risk with risk differs from—as well as shares affinities with some elements of—the two sides of the dichotomous precautionary and innovation-oriented paradigms that make up contemporary debate over risk governance.

We offer a framework for deciding when government is justified to fight risk with risk, and when it should revise or discontinue such efforts. This inquiry implicates first-order questions—namely, the welfare implications of risk–risk tradeoffs and the distributive consequences of shifting exposures to risks—and second-order questions concerning the epistemic challenges in authorizing or deploying risk as a strategy for risk governance. These challenges involve making decisions under uncertainty across the policy life cycle, mitigating institutional and behavioral biases, structuring liability and compensation regimes for the risks that law deliberately embraces, and building reflexivity in an often-fragmented institutional environment.

To address these challenges, we advance a meta-regulatory architecture rooted in reflexivity—the capacity of legal institutions to adapt and update policies as evidence, risks, and priorities evolve. This architecture rests on four interlocking elements, each of which corresponds with the main challenges we identify: dynamic updating; structures to counteract cognitive biases; targeted liability and compensation; and collaborative mechanisms that enable coordinated revision. Without such governance mechanisms, regulatory authorities that embrace risk may inadvertently entrench policies whose harms outweigh their benefits and whose inequitable effects deepen existing vulnerabilities in society.

By highlighting how government can affirmatively embrace risk as a risk governance strategy, we seek to bring a fresh perspective to the precaution-versus-innovation debate and point to other ways that law can confront the harms facing society today. The debate between precaution and innovation is incomplete, if not also overly simplistic. Regulating in a highly dynamic world necessitates an equally dynamic posture toward risk. In a fast-paced era defined by climate change and new technologies, society surely requires some precautionary forms of risk governance that minimize risk, as well as some government policies that promote innovation. But it will also at times require strategies that deliberately deploy risk to avert greater harm. The meta-regulatory framework we propose offers a blueprint for embracing risk by integrating economic efficiency with a commitment to distributive fairness. That framework also points to how regulators should act more generally in the face of the risks of new technologies and a changing world.

Paul Aubrecht (U Passau) and Mitja Kovac (U Ljubljana) have posted “AI Regulation: A Comparison of Centralized and Decentralized Approaches in Federal Systems” on SSRN. Here is the abstract:

The regulation of artificial intelligence (AI) is at the forefront of legal scholarship, as its use has the potential to change nearly every aspect of human society. The potential for AI to impact society makes its regulation particularly relevant to legal scholars, practitioners, and lawmakers.

This research considers the benefits of centralization and decentralization in regulating AI and the implications of convergence or divergence in AI regulation within a federal system. We examine the United States (US) and European Union (EU) approaches to regulation of AI and show that there is a distinct difference in risk preferences related to the regulation of AI between the EU and US as well as a higher likelihood of negative externalities related to the regulation of AI emerging in the US than in the EU while also recognizing that the US approach of decentralized regulation of AI is more likely to lead to the identification of novel approaches to the regulation of AI which may eventually lead to positive externalities. Thus, we generally consider that each approach is not risk-neutral, though for very different reasons.

This examination focuses on the costs and benefits of centralization and decentralization within federal systems, i.e., systems in which regulatory competencies are divided between the federal and state levels. The US and EU provide useful examples of how federal approaches to regulating AI diverge. A quick look at the regulation of AI in the EU and the US shows a divergence in approaches to the regulation of AI within borders, between states inside federal systems (within the US), and across borders between different federal systems (EU and US).

Lucia Velasco (Maastricht U Business and Economics) et al. have posted “The Future of the AI Summit Series” on SSRN. Here is the abstract:

The AI Summit series – initiated at Bletchley Park in 2023 and continuing through Seoul in 2024 and Paris in 2025 – has become a distinct forum for international collaboration on AI governance. Its early achievements, including the Bletchley Declaration, the Frontier AI Safety Commitments, and the International Scientific Report on the Safety of Advanced AI, are a result of its unique format, regular schedule, and ability to secure concrete commitments from governments and industry.

To ensure its continuing impact, the Summit series must now transition from an improvised sequence of summits towards a more formalized structure. For this evolution to succeed, organizers must carefully examine past successes and realistically assess future challenges. This report examines both, with particular attention to a set of core summit design elements: hosting arrangement, secretariat format, participant selection, agenda setting, and summit frequency. Based on this analysis, we present six recommendations to strengthen the summit series’ impact.

The paper draws on existing international governance models to offer recommendations for each design element, addressing challenges such as a crowded summit landscape, geopolitical shifts, and rapid technological change.

Caitlin Burke (Stanford U) has posted “TikTok, Instagram, and the “Fourth Party”: The Impact of Technical Design on Personal Content Moderation” (Ohio State Journal On Dispute Resolution 2025) on SSRN. Here is the abstract:

As policymakers debate Section 230, children’s online safety, andsocial media regulation, increasing attention has shifted from online content to the design of digital platforms. This article examines how thepersonal content moderation systems of TikTok and Instagram shape user experiences through interface design, reporting tools, appeals processes, and platform governance. Drawing ondispute system design, it introduces the concept of the “fourth party” to explain how technical design influences online disputes and redistributes power between platforms and users. The article argues thatplatform design is analytically distinct from protected speech and should therefore be evaluated separately from content moderation under theFirst Amendment. By reframing content moderation as a problem of product and interface design rather than speech alone, the article offers a framework forplatform accountability, consumer protection, and online safety that preserves free expression while addressing the harms of modern social media.

Urooba Salman (Lahore U Management Sciences (LUMS)) has posted “The Digital Panopticon: How AI has Reshaped Workplace Surveillance and Labor Control” on SSRN. Here is the abstract:

This paper seeks to interrogate how artificial intelligence has transformed the exercise of managerial power, labor autonomy, and the traditional and foundational principles of labor law. Drawing on Michel Foucault’s theory of thepanopticon and extending it to the modern workplace, it argues that algorithmic systems have created a “digital panopticon”; an architecture of invisible surveillance which disciplines workers through datafication and the fear of surveillance rather than outright coercion. Using insights from labor law scholarship, critical theory, and existing empirical research, the paper demonstrates that AI-based monitoring tools, ranging from productivity trackers and biometric attendance systems to predictive analytics, actually embed control into the very scaffolds of the modern workplace, eroding collective bargaining practices, privacy, and procedural fairness.

This study situates these changes and transformations within the Pakistani context and legal framework, where though the infiltration of surveillance is not so perverse, fragile labor protections, weak privacy and data governance laws, and imported surveillance technologies create an especially fertile ground for exploitation. It conceptualises this phenomenon as “surveillance colonialism,” wherein technologies from authoritarian or unregulated jurisdictions infiltrate developing economies under the guise of modernization. The paper contends that Pakistan’s outdated legal architecture regarding AI and data cannot adequately address such algorithmic decision making, leading to an accountability vacuum and a structural silencing of labor resistance which needs to be addressed before it is too late.

Tom Golway (Generative Dynamics Research) has posted “The Problems of Philosophy in the Age of AI” on SSRN. Here is the abstract:

When Bertrand Russell wrote The Problems of Philosophy in 1912, he grappled with the gap between appearance and reality, asking how we can know anything with certainty when our senses may deceive us. Russell’s skepticism presumed that underlying truth existed and could be approached through rigorous inquiry. Over a century later, his questions have not merely persisted—they have proliferated into new domains of epistemic risk. Artificial intelligence does not simply introduce fresh uncertainties; it actively manufactures realities, fragments shared understanding, and operates at speeds that preclude human deliberation. In this landscape, the peril is not ignorance but epistemic surrender: the quiet abdication of judgment to systems that neither know nor care what is true. This paper revisits Russell’s inquiry in light of AI’s epistemic power, arguing for a renewed ethics of validation, provenance, and human oversight.

Valerio De Stefano (York U Osgoode Hall Law) has posted “Labour Law, Technology, and the Attack on the Rules-Based-Order” on SSRN. Here is the abstract:

This paper argues that current debates on artificial intelligence and algorithmic management at work are also debates about private power, managerial prerogative, and democracy. It situates recent attacks by major technology companies and their political allies on the European Union and other rule-based international institutions within a broader struggle over who should regulate technology and govern society. It contends that the anti-democratic turn of parts of the tech world is closely connected to authoritarian ideas about work, hierarchy, and obedience. The workplace is one of the principal sites where authority is exercised, surveillance normalised, and habits of subordination formed, with consequences that extend beyond employment relations into democratic life more broadly. Against this background, the paper examines how AI and algorithmic management intensify employer power through pervasive monitoring, automated evaluation, predictive analytics, and data-driven discipline. It argues that existing legal frameworks, especially data protection law, remain insufficient unless infused with labour law concepts that address the structural imbalance of power at work. It therefore advocates a substantial upgrading of labour law, stronger limits on digital surveillance, and a more robust role for collective bargaining, codetermination, and worker voice in the governance of workplace technologies. It contends that labour law is one of the main ways in which the rule of law enters the workplace and workers remain citizens rather than subjects while at work. The regulation of AI at work is thus a constitutional and democratic question as much as an economic or technological one.

Joshua S. Joseph (Hofstra U Maurice A. Deane Law) has posted “Balancing Innovation and Biomedical Ethics within National Institutes of Health: Integrative and Regulatory Reforms for Artificial Intelligence-Driven Biotechnology” (44 Biotech. L. Rep. 2 (2025) [10.1089/blr.2025.360002.jj ]) on SSRN. Here is the abstract:

The article examines the foundational principles of biomedical ethics and their relevance to the intersection of artificial intelligence (AI) and biotechnological progress. It then analyzes the potential impact of the National Institutes of Health’s (NIH or the Agency) proposed 2024 reorganization reforms, as presented by the House Committee on Energy and Commerce (E&C), and advocates for the inclusion of provisions that would enable the Agency to develop and deploy “safe, secure, and trustworthy” AI within the biotechnology sector. Thus, the article proposes two key integrative reforms to address these concerns: (1) enhancing research collaborations among healthcare systems, academic institutions, and the private sector through AI-driven platforms, and (2) establishing an NIH office to oversee AI-based health initiatives. These reforms would require legislative amendments, including updates to the Patent and Trademark Law Amendments Act of 1980 (Bayh-Dole Act or Bayh Dole) and the 21st Century Cures Act of 2016 (Cures Act), while adhering to biomedical ethics. Such amendments would ensure (1) ethical standards in collaborative research and development (R&D) among institutions and (2) effective health initiative design and implementation. In the same light, this article advocates for an updated legislative framework to the Health Insurance Portability and Accountability Act of 1996 (HIPAA) that regulates the role of such integrative efforts and balances the transformative potential of AI in biotechnology. In all, these regulatory steps ensure that innovation benefits society while safeguarding individual rights such as personal data and privacy and mitigating algorithmic bias.

Frank Pasquale (Cornell U Law) has posted “Discerning Artificial and Authentic Intelligence: Profundity as Depth in Antiqua et Nova” on SSRN. Here is the abstract:

The rise of artificial intelligence challenges institutions to help persons chart a path toward authentic intelligence. There are both objective (the real) and subjective (the sincere) dimensions of authenticity. By advancing awareness of reality and cultivation of sincerity as the foundations of authentic intelligence, Antiqua et Nova helps us move beyond the fetishization of artifacts at the core of TESCREAL and AGI ideologies, toward profound insight into the unique value of the human person. Cognitive virtues of seriousness, grounding, and wisdom, rooted in human experience, all support deep thought. Antiqua et Nova (AN) offers a philosophical anthropology to overcome the shallowness (and eventual surrender) of subjectivity portended by a gradual erasure of boundaries between persons and machines. An incarnational philosophy of mind and body informs Antiqua et Nova’s understanding of intelligence, motivating critical distinctions between authenticity and artificiality.

Benjamin Minhao Chen (The U Hong Kong Law) and Xinyu Xie (The U Hong Kong Law) have posted “The Alignment Target Problem: Divergent Moral Judgments of Humans, AI Systems, and Their Designers” on SSRN. Here is the abstract:

The project of aligning machine behavior with human values raises a basic problem: whose moral expectations should guide AI decision-making? Much alignment research assumes that the appropriate benchmark is how humans themselves would act in a given situation. Studies of agent-type value forks challenge this assumption by showing that people do not always judge humans and AI systems identically.This paper extends that challenge by examining two further possibilities: first, that evaluations of AI behavior change when its human origins are made visible; and second, that people judge the humans who program AI systems differently from either the machines or the human actors they are compared against. An experiment with 1,002 U.S. adults measured moral judgments in a runaway mine train scenario, varying the subject of evaluation across four conditions: a repairman, a repair robot, a repair robot programmed by company engineers, and company engineers programming a repair robot. We find no significant difference in evaluations of the repairman and the robot. However, judgments shifted substantially when the robot’s actions were described as the product of human design. Participants exhibited markedly more deontological, rule-based reasoning when evaluating either the programmed robot or the engineers who programmed it, suggesting that rendering human agency visible activates heightened moral constraints. These findings indicate that people may evaluate humans, AI systems acting in the same situation, and the humans who design them in meaningfully different ways. The fact that these evaluations do not necessarily converge gives rise to the alignment target problem: which normative target should guide the development of artificial moral agents in high-stakes domains, and whether these plural judgments can be reconciled within a coherent account of value alignment.