About | aiD

The Norwegian Centre on AI for Decisions (aiD) is a premier research hub dedicated to advancing the role of artificial intelligence in complex decision-making processes. As a cornerstone of the Research Council of Norway's (RCN) AI portfolio, aiD bridges technological, organizational, and human-centric gaps to foster a society where AI-driven value creation is safe and ethical.

What is aiD?

Academic Institutions

Research Organizations

Professional Organizations (POs)

aiD is an interdisciplinary collaboration involving 3 academic institutions, 6 research organizations, and over 50 professional organizations (POs). The center is led by Director Prof. Sebastien Gros (NTNU) and Co-Director Dr. Signe Riemer-Sørensen (SINTEF Digital), supported by a management team with responsibility for fundamental research, operationalization, and innovation.

Our primary objective is to advance AI for decision-making through fundamental research and real-world use cases, ensuring that AI-enhanced human decisions or fully autonomous systems are effective, safe, and trustworthy in sectors critical to society.

aiD moves beyond traditional AI models that are purely descriptive. Instead, we employ a holistic framework where the primary learning objective is the decision outcome itself. This involves backpropagating learning through the entire decision process, from converting raw data into information to predicting how actions impact the real world while modeling risk and human preferences.

Our research addresses several critical challenges that currently limit AI adoption in professional organizations, including low trustworthiness, immature technology, and a shortage of expertise regarding emerging regulations.

Research areas

To achieve its goals, aiD operates across nine specialized Research Areas (RAs):

Trust

Developing frameworks for safe, responsible AI and robust system verification, assurance and compliance.

Barriers

Solutions for technological and organisational barriers for AI in decisions processes.

Knowledge-Embedding

Integrating physical laws, simulations, and domain expertise into AI via Scientific Machine Learning (SciML) to ensure reliability even with scarce data. Machine-learning driven causal inference in decision processes.

Data to Information

Finding optimal latent data representations with maximum decision relevant information.

Training

Creating novel loss functions and training methods tailored to decision objectives like safety and utility rather than generic accuracy.

Generalization and Uncertainty

Managing how AI models extrapolate beyond their training data to prevent unsafe decisions in uncertain environments.

Human-AI Interaction

Designing interfaces that combine AI's speed with human intuition, ensuring operators can trust and understand AI-supported decisions.

Multi-Agents

Developing fair and transparent frameworks for scenarios where multiple AI or human agents share resources or have competing objectives.

Disruptive

Monitor other technologies that will influence AI and adapt to the development.

Real-World Impact and Use Cases

aiD demonstrates its research progress through more than 10 high-impact use cases organized into five topical areas and driven by industrial partners:

Real-World Impact and Use Cases

aiD demonstrates its research progress through more than 10 high-impact use cases organized into five topical areas and driven by industrial partners:

Risk-based maintenance

Predicting failures and recommending actions for wind turbines, power grids, ships, and buildings.

Autonomy

Removing humans from repetitive or dangerous processes in maritime navigation and process control.

Resource and Asset Optimization

Improving patient treatment pathways in healthcare and energy management in large building portfolios.

Operations Advancement

Enhancing product quality and streamlining business operations in the process and maritime industries.

Logistics and Supply

Optimizing delivery management, passenger transportation, and road maintenance.

Mission and Goals

A core mission of AID is to develop the next generation of AI experts. We aim to educate 400 MSc students and 30 PhDs/postdocs, while strengthening the expertise of over 50 existing researchers. By pairing academic researchers with professional partners, we ensure that our training is grounded in real-world industrial and societal needs.

Key milestones

400 MSc students educated

30 PhDs/postdocs graduated

50 Researchers strengthened

Sustainablity and Ethics

Our research is fundamentally aligned with the UN Sustainable Development Goals (SDGs), specifically contributing to affordable and clean energy (SDG 7), good health and well-being (SDG 3), and sustainable industry and infrastructure (SDG 9). We strictly adhere to principles of Fairness, Accountability, Transparency, and Explainability (FATE) to ensure our solutions comply with the EU AI Act and OECD recommendations for trustworthy AI.