Recommender systems can play a significant role in supporting decisions in environments characterized by overwhelming choice – such as education and professional training. While Aritifical Intelligence (AI) can provide humans with highly personalized and accurate recommendations, such decisions have long-term implications for personal and professional development and thus demand a high degree of reflection.

In several projects, we develop AI as a reflection partner for lifelong learning. Based on Explainable AI (XAI), recommendations are made transparent: users can see which skills, interests, and goals drive AI recommendations. They are able to adjust priorities, explore alternatives, and simulate different learning scenarios. This creates a co-creative process in which humans and AI jointly explore potential educational pathways.

The concept has been implemented and investigated in several projects, for instance being “Kris Känguru” in collaboration with the START Foundation to support students with migration backgrounds, “XPERT” for professional training contexts,“AI Scouty” to assist students in course selection, and is currently being explored together with Senior Connect.

Findings demonstrate that combining recommendation systems with explainability and interactivity enhances user engagement and decision quality. Rather than passively accepting recommendations, users actively interpret, question, and refine AI recommendations. The system continuously learns from user interactions, leading to increasingly personalized and context-aware recommendations over time. This approach shifts the role of AI from a static advisor to a dynamic, adaptive partner in decision-making.

The concept is highly transferable, with applications ranging from career guidance for young people to workforce development in organizations. In the context of rapid technological change and evolving skill requirements, the approach provides a scalable solution to support lifelong learning.