Artificial Intelligence (AI) at SnT
Our scientific programme

How we approach AI research

SnT’s AI research is organised around four complementary themes, covering the full spectrum from foundational methods to applied systems and societal impact. Each theme is a scientific programme, not just a topic area: it has open questions, methods, obstacles, target environments, and expected outputs.

Four AI research themes

Scientific questions & methods

How do we make AI models explainable without sacrificing performance? How do we evaluate AI against regulatory standards such as the EU AI Act? Methods include formal verification, adversarial testing, counterfactual explanation, and statistical fairness analysis.

Obstacles & target environments

Key obstacles are the gap between model internals and human-interpretable explanations, and the absence of agreed evaluation benchmarks. Target environments include regulated sectors – financial services, healthcare, public administration – where opacity is a compliance risk.

Expected outputs

Evaluation frameworks, auditing tools, bias benchmarks, normative contributions to EU AI governance, validated explainability methods, and NLP tools for automated compliance reasoning.

AI research projects in action

These three projects illustrate how SnT works – demanding sectors, real deployment conditions, and identifiable outputs.

AI as a research accelerator

Groups developing AI itself

These groups advance the methods, architectures, and theory of AI, building the foundational tools that others can apply and deploy.

Groups using AI as a scientific accelerator

These groups apply AI to drive breakthroughs in their own domain – from legal reasoning to health monitoring – generating new use cases that feed back into AI development.

Scientific outputs

Interested in our research ?

Share this