Open UnP Questions in AI Regulation

As decision-making becomes increasingly distributed across human collectives and technological systems, the UnP tension takes new forms:

  • How can hybrid intelligence models balance rule-based reasoning with contextual understanding?

  • How do we program flexibility into algorithmic systems?

  • And how can institutions leverage both universal standards and particular insights to navigate complex trade-offs between competing values like accountability and forgiveness?

This integrated challenge invites us to move beyond convenient dichotomies toward more nuanced approaches that recognize the necessary dialogue between principles and particulars across all domains of human thought and action.

While the institute’s research is focused on UnP problems as basic science questions, our aim is to use this knowledge within the space of regulatory mechanisms — what good is knowing about rules if you don’t craft better rules?

The questions are split into three main hubs (I really like AI generated acronyms)

COSMIC (Cognitive Systems and Mechanisms In Information Ecosystems):

NOMOS (Norm Ontogeny, Maintenance, Outflow, and Systems)

MOSAIC (Mechanisms Of Socio-Legal AI Compliance)

    • How do individuals and AI systems mentally model and navigate complex information landscapes in order to make decisions?

    • To what extent does prior knowledge influence the interpretation of new information in specialized informational contexts (e.g., scientific literature, legal documents)?

    • To what extent does framing (personas, prompt engineering, collective identity, etc) affect the interpretation of information and decision justification/explanation, and why?

    • How do various agents (human, simulated/artificial, and hybrid) interact in the process of norm creation and maintenance?

    • How does the authorship of formal rules (laws, regulations) and informal norms shape decision making behaviour of different agents

  • Connecting to COSMIC:

    • How do AI systems and human policymakers mentally model complex legal and regulatory landscapes, and how does this affect governance decisions?

    • What cognitive mechanisms are involved in interpreting and applying legal frameworks to novel AI-related scenarios?

    • How does framing (e.g., anthropomorphising of AI, technical jargon, ethical considerations, hype) influence the development and interpretation of AI governance policies

    Connecting to NOMOS:

    • - How do norms emerge and evolve in the interaction between AI systems, human agents, and legal frameworks?

    • What role do AI systems play in shaping, interpreting, and enforcing legal norms in various contexts?

    • How does the authorship of AI-related regulations (e.g., by technologists, lawyers, or hybrid teams) affect their interpretation and implementation by different agents?

    Bridging COSMIC and NOMOS:

    • How can we design legal frameworks that account for both human and AI cognitive mechanisms in information processing and decision-making?

    • What are the cognitive and normative challenges in creating governance structures that remain effective as AI capabilities evolve?

    • How do cultural differences in information processing and norm adherence affect the global governance of AI systems?