Generative AI-supported Worker Training in Virtual Reality
Project Description
Many industries continue to face significant challenges in workforce training regarding safety and skill acquisition. Despite technological advancements, traditional training methods remain largely static and classroom-based. They are also highly costly, as instructors can only provide personalized feed-back in small groups, often only in 1:1 training sessions. One hour of (personalized) employee training can cost around 100 CHF. Thus, traditional methods typically fail to provide immersive, adaptive learning experiences that can effectively prepare workers for the complex and often dangerous envi-ronments they encounter, for instance, on construction sites.
Virtual reality (VR) is well-known for effective worker training. However, the setup and implementation of new products in VR training are costly and time-consuming. This project aims to tackle these chal-lenges using artificial intelligence supporting employees in VR training. In particular, we seek to lever-age state-of-the-art generative AI to access existing knowledge in text documents, such as technical documentation or product descriptions, and make the knowledge accessible during VR training. More-over, we aim to apply generative AI to assess training sessions to provide immediate feedback to trainees. Additionally, as documentation of 3D models is often sparse with non-standardized naming, we aim to use AI to link and visualize textual information better and automatically.
Our approach will provide a practical contribution in the form of a highly innovative AI-based solution for the problem at hand and a scientific contribution reporting the results of the proposed design science research approach.
Virtual reality (VR) is well-known for effective worker training. However, the setup and implementation of new products in VR training are costly and time-consuming. This project aims to tackle these chal-lenges using artificial intelligence supporting employees in VR training. In particular, we seek to lever-age state-of-the-art generative AI to access existing knowledge in text documents, such as technical documentation or product descriptions, and make the knowledge accessible during VR training. More-over, we aim to apply generative AI to assess training sessions to provide immediate feedback to trainees. Additionally, as documentation of 3D models is often sparse with non-standardized naming, we aim to use AI to link and visualize textual information better and automatically.
Our approach will provide a practical contribution in the form of a highly innovative AI-based solution for the problem at hand and a scientific contribution reporting the results of the proposed design science research approach.
