The OptiPEx project aims to improve the passenger experience of public transport, by enhancing their sense of comfort, safety, and ease of travel. This increase in service quality and improvement of passenger satisfaction will lead to an acceleration of the adoption of automated public transportation technologies, and ultimately improve the sustainability of the mobility services across Europe. As a part of this, OptiPEx has studied the factors that support or constrain the development of passenger-aware AI services in public transport. In doing so, the project is developing a structured analysis framework, at the service concept level, that identifies enablers and barriers across multiple dimensions and supports transferable insights beyond the OptiPEx project. The data for this analysis included framework and standards reviews; service concept materials; expert workshop analysis and results; and, service concept evaluation results from surveys and focus groups. In summary, OptiPEx aims to identify the key enablers and barriers to the design and deployment of passenger-aware AI services in public transport, using service concepts—concrete representations of proposed passenger-facing services and their intended functions—as the unit of analysis.
For each service concept, the analysis considered the passenger-related data involved, the applicable legal and ethical requirements, potential safety and security implications, evidence needed for technical validation, and operational arrangements required for real-world deployment. The findings are being documented in an enabler–barrier matrix linking each service concept to the five aforementioned analysis dimensions within the five following cross-cutting enablers and barriers: privacy-preserving design; representative data; technical robustness; organisational readiness; passenger trust; and human oversight. While the full matrix will be made available soon, this article offers a snapshot of these findings pertaining to the issue of human oversight. These perceptions are strongly indicated in the convergence of two of the key quantitative data sources used: the surveys and the focus groups. The findings can be broken down into the seven key perceptions below.
- Human oversight as a source of safety
Passengers associate drivers and onboard staff with safety and supervision. Their presence provides reassurance that someone can respond to accidents, technical failures, emergencies or unexpected situations. Fully driverless transport fosters a sense of uncertainty around who will intervene when something goes wrong. - Human oversight as a form of assistance and accessibility
Although participants recognise the potential of automation to improve accessibility through features such as automatic ramps, doors and passenger detection, they still value human assistance. This is particularly important for older passengers, wheelchair users and people with visual or reduced mobility, who may require help when automated systems do not work as expected - Human oversight as responsiveness to passenger needs
Participants question whether an autonomous system can adequately understand and respond to individual or unexpected passenger needs. Human staff are perceived as having the ability to interpret situations and adapt their response, whereas automated systems may be limited to predefined or detected conditions. - Human oversight for conflict and emergency management
Participants are generally comfortable with AI being used to detect altercations, unusual events or emergencies, but are reluctant to give AI the responsibility for deciding how these situations should be managed. Human staff are seen as necessary to interpret complex situations and determine an appropriate response - Human oversight as reassurance and trust
The presence of a driver or onboard staff member provides passengers with a sense that someone is responsible for the service and their wellbeing. This human presence contributes to feelings of reassurance and trust, particularly when passengers are uncertain about how an autonomous system would respond to an unexpected situation - Human oversight as management of shared spaces
Passengers perceive human staff as important for maintaining order within the vehicle. They can manage reserved seating, address inappropriate or uncivil behaviour, and assist passengers who need additional support. Although cameras and intelligent systems can monitor these situations, participants still see value in having a person who can act on this information - Human oversight as a complement to AI
Participants are generally receptive to technologies such as cameras, sensors, automatic alerts, collision-prevention systems and automated accessibility functions when these technologies support human staff. AI is therefore viewed positively when it provides an additional layer of information or safety rather than assuming complete responsibility
The perceptions summarised above point to a clear distinction between perceptions regarding automation as support and automation as human replacement. More specifically, the data shows that participants generally accept AI and automated technologies that can improve safety, accessibility, monitoring, and information provision when it is used to support the efforts of human staff, while are notably more cautious about fully autonomous public transport where no human is available to supervise, assist or intervene. As such, findings indicate that participants’ concerns are primary about automation that involves the complete removal of human staff, and not automation itself.
This means that human oversight remains an important condition for trust and acceptance of autonomous public transport, and is a key insight that OptiPEx will use to guide the development of our services and the technologies that support them.
