Inclusive AI for Mobility and Transportation. Oral Presentation at the 17th NNDR Conference: Disability in Local and Global Contexts. 7-9 May 2025, University of Helsinki, Finland

Övrigt konferensbidrag, 2025

Background aims: During the years 2024-2026, a multidisciplinary research project (No AI About Us Without Us: Enhancing Safety for Vulnerable Road Users (VRUs) with Reduced Mobility" and abbreviated to NoAI) is underway at University of Gothenburg and Chalmers University of Technology in Sweden about using practical knowledge and experiences in people with various disabilities (visual impairment and blindness, hearing impairment, motor impairment) to create safer mobility for all in society. The project aims to identify gaps in current road infrastructure and develop an AI-driven system to enhance accessible urban planning, while, in the future, also warning and assisting VRUs in potentially dangerous traffic situations and risks. Methods: The study includes researchers in social work (disability research), computer science, and design. Both qualitative and quantitative research methods have been used. Interviews, observations, filming, diary writing, and AI-driven analyses have been carried out in the project. During the spring and autumn, 20 people with different disabilities were interviewed about their experiences of mobility and movement in external environments. Observations and filming of the participants' movement patterns in different traffic environments and risk situations were carried out during September-October 2024. Results and conclusions: Our study reveals experiences of insecurity, risks, incidents and accidents emerged, but also the need for various safety-creating measures. In particular, we found that the design of heterogeneous road traffic control devices must be more carefully examined to ensure their effectiveness, preventing people in need from having to find their own workarounds. Furthermore, the diversity in motion patterns and appearances among VRUs, particularly those with disabilities or using mobility aids, complicates their identification by both other road users and AI systems. This issue is compounded by the increased risk of collisions faced by these individuals. These findings, highlighting these challenges, will be presented at the conference.