Skip to main content
Aller à la page d’accueil de la Commission européenne (s’ouvre dans une nouvelle fenêtre)
français fr
CORDIS - Résultats de la recherche de l’UE
CORDIS

Governing Generative AI from the Inside and Out: The Politics and Practices of AI 'Alignment'

Objectif

‘Generative’ artificial intelligence tools, like large language model-powered chatbots, are promising to transform how people access information and interact with the social and political world. This has suddenly made the technology companies that build AI models (and the interfaces, platforms, filters, and guardrails that mediate access to those models) a globally salient site of political contestation from individuals, civil society, and government actors seeking not just to shape how these systems work, but also to harness them for their particular preferences and needs. ALIGNING-AI is an empirical project examining the complex politics emerging at this particular intersection of private rulemaking, sociotechnical system design, and global interest group politics. The project investigates how AI development actors respond to diverse stakeholder demands and encode them — or not — into the systems they build, taking a cross-national perspective and employing methodological approaches drawn from both political science as well as digital media studies and human-computer interaction. Combining deep interviews with AI developers with longitudinal multi-platform data on model performance across various jurisdictions, the project will explain how actors inside AI firms and outside of them govern generative AI technologies, exploring variation across different firms, countries, and AI development modalities.

ALIGNING-AI will thus offer the first detailed analysis of the governance practices of AI ‘alignment,’ map the emerging global space of stakeholder contestation around genAI models, create open datasets to be used by other researchers interested in the technical and political practices of AI alignment, and outline policy recommendations on how to make these practices more accountable going forward.

Champ scientifique (EuroSciVoc)

Ce projet n'a pas encore été classé par EuroSciVoc.
Proposez les domaines scientifiques qui vous semblent les plus pertinents et aidez-nous à améliorer notre service de classification.

Mots‑clés

Programme(s)

Thème(s)

Régime de financement

HORIZON-ERC - HORIZON ERC Grants

Voir tous les projets financés dans le cadre de ce programme de financement

Appel à propositions

(s’ouvre dans une nouvelle fenêtre) ERC-2026-STG

Voir tous les projets financés au titre de cet appel

Institution d’accueil

WISSENSCHAFTSZENTRUM BERLIN FUR SOZIALFORSCHUNG GGMBH
Contribution nette de l'UE
€ 1 499 875,00
Coût total
€ 1 499 875,00

Bénéficiaires (1)