State of the Field
Recent research in AI governance is increasingly focused on establishing frameworks that can keep pace with the rapid evolution of AI technologies. One significant area of exploration is the automation of AI research and development, which raises questions about the balance between capability advancement and safety oversight. Metrics have been proposed to better understand this dynamic, enabling stakeholders to track the implications of AI R&D automation. Concurrently, the introduction of the Sentience Readiness Index highlights the need for national preparedness in the face of potential AI sentience, revealing widespread inadequacies in institutional and cultural readiness. Additionally, the concept of Institutional AI is gaining traction, proposing governance structures that can mitigate risks associated with multi-agent AI systems. As AI continues to permeate various sectors, these developments underscore the urgency for robust legal and regulatory infrastructures that not only set rules but also adapt to the complexities of AI decision-making, ensuring that human oversight remains integral in an increasingly automated landscape.
Papers
1–10 of 12Measuring AI R&D Automation
The automation of AI R&D (AIRDA) could have significant implications, but its extent and ultimate effects remain uncertain. We need empirical data to resolve these uncertainties, but existing data (pr...
The Sentience Readiness Index: Measuring National Preparedness for the Possibility of Artificial Sentience
The scientific study of consciousness has begun to generate testable predictions about artificial systems. A landmark collaborative assessment evaluated current AI architectures against six leading th...
From Reflection to Repair: A Scoping Review of Dataset Documentation Tools
Dataset documentation is widely recognized as essential for the responsible development of automated systems. Despite growing efforts to support documentation through different kinds of artifacts, lit...
Institutional AI: Governing LLM Collusion in Multi-Agent Cournot Markets via Public Governance Graphs
Multi-agent LLM ensembles can converge on coordinated, socially harmful equilibria. This paper advances an experimental framework for evaluating Institutional AI, our system-level approach to AI align...
AI Narrative Breakdown. A Critical Assessment of Power and Promise
This article sets off for an exploration of the still evolving discourse surrounding artificial intelligence (AI) in the wake of the release of ChatGPT. It scrutinizes the pervasive narratives that ar...
Making Models Unmergeable via Scaling-Sensitive Loss Landscape
The rise of model hubs has made it easier to access reusable model components, making model merging a practical tool for combining capabilities. Yet, this modularity also creates a \emph{governance ga...
Explicit Cognitive Allocation: A Principle for Governed and Auditable Inference in Large Language Models
The rapid adoption of large language models (LLMs) has enabled new forms of AI-assisted reasoning across scientific, technical, and organizational domains. However, prevailing modes of LLM use remain ...
Architecting Trust in Artificial Epistemic Agents
Large language models increasingly function as epistemic agents -- entities that can 1) autonomously pursue epistemic goals and 2) actively shape our shared knowledge environment. They curate the info...
Legal Infrastructure for Transformative AI Governance
Most of our AI governance efforts focus on substance: what rules do we want in place? What limits or checks do we want to impose on AI development and deployment? But a key role for law is not only to...
Delegation Without Living Governance
Most governance frameworks assume that rules can be defined in advance, systems can be engineered to comply, and accountability can be applied after outcomes occur. This model worked when machines rep...