In the “Uni Bremen informs” series, we report on the research and development activities of the University of Bremen on the topic of AI in education (or related to it) and refer to central project results.

In the logistics sector, we see increasing substitution risks in addition to advancing digitalization and automation. There is also the challenge of counteracting the ongoing shortage of skilled workers. In addition to the quantitative shortage of skilled workers, there is also a need for qualitatively trained employees. However, employee participation in further training is rather low. As undisputed as the need for learning is, it is difficult to develop further training formats in such a way that they address both company and professional perspectives.

The SMALO “Smart Learning in Logistics” project (project duration: September 1, 2021 to August 31, 2024) addresses the aforementioned challenges by designing and developing platform-based training for employees in logistics in the sense of a digital learning ecosystem.

The aim of the project is to develop a demand- and application-oriented digital and secure learning space that supports networking activities at an individual and institutional level using the example of planning and operational fields of action in logistics. The different perspectives and needs of users are identified and learning paths are provided that enable individual and flexible learning. The design and development of the SMALO learning ecosystem goes beyond the mere networking of learning locations by increasing the transparency of suitable training opportunities. At the personnel level, this can increase the matching of suitable training opportunities and at the same time enable resource-oriented training and further education at company level.

The AI-based services in the SMALO learning ecosystem are implemented on a dedicated server and called up by the SMALO system via a REST API (REST = Representational State Transfer, API = Application Programming Interface). The AI solutions are implemented in Python due to its powerful libraries for neural networks such as TensorFlow and Keras. Data analytics are used to categorize course content for targeted suggestion generation in the database and for the AI-supported creation of learning paths based on user profiles. Learning support is provided on SMALO by the chat bot “Lisa-Logistic”. Lisa-Logistic welcomes the learner and uses onboarding questions to select whether adaptation training or rather advanced training is an option based on individual needs. The learner is then offered suitable training courses from external training service providers, which can be freely selected in face-to-face or online courses depending on the learner’s preference. A learning path tailored to the learner is then generated on the basis of the selected courses and saved under the “My learning path” function.