Reimagining Values as Cross-Level Intelligence

Concept image of an AI wearing a digitized human face

Project description

This project investigates values as a form of cross-level intelligence emerging from interaction rather than residing solely within individuals. Challenging individualist models, it examines how values organise intelligent behaviour and how they develop across scales.

Primary participants

Principal Investigator:

Dr Vladimir Ponizovskiy, Psychology
vladimir.ponizovskiy@durham.ac.uk

Term:

Epiphany and Easter 2027

The project will integrate perspectives from psychology, computer science, economics, philosophy, anthropology, and the arts. By creating new interdisciplinary synergies, the project aims to open provocative perspectives on intelligence, with implications for understanding learning and decision-making, social coordination and collective action, cultural change and adaptation, and the alignment of artificial systems with human values.

Return to all 'Research Development Projects'

Overview:

The project takes as a starting point a critique of a dominant assumption in European thought, following Descartes, that mind and intelligence are located primarily within an individual. This assumption continues to shape psychology’s focus on intra-individual processes, the design of frontier AI agents as solitary containers of knowledge, and cultural forms that focus on individual experience. Across disciplines, however, there is growing dissatisfaction with this view. Intelligence is increasingly understood as emerging from interaction: feedback loops, coordination, and dynamic relations between agents and their environments. Particularly promising in this respect is the construct of values (desirable end-goals, principles, preferences, and negotiated worth) which emerge between individuals but organise individual priorities and action. Through workshops and reading groups, the PI and participants will examine and critique disciplinary-specific conceptualisations of intelligence and values. This work will explicitly challenge Cartesian individualism and foreground interaction: feedback loops, coordination, dynamic relations between agents and their environment. The goal of this stage is not synthesis but the development of shared conceptual language across disciplines. Building on the exchange, a multi-level framework will be developed that links (1) individual learning and goal-setting, (2) shared norms and coherent collective behaviour, and (3) longer-term cultural dynamics. Particular attention will be directed at feedback loops between levels, interaction between systems and their environment, and value negotiation, stabilisation, and change in human and artificial systems. In the final stage, the team will identify high-impact research questions and methodological approaches (experimental, computation, ethnographic, artistic) suitable for a large-scale interdisciplinary project. The IAS RDP will culminate in a joint grant application (e.g. Leverhulme, ERC) that would advance this important frontier.