The online meeting was structured around two thematic sections: AI supporting public transport, and AI supporting transport beyond mass transit and city limits.

AI supporting public transport: From promise to practice

The session opened with concrete uses of AI already visible in operations: forecasting passenger flows, adjusting service frequency and routes in real time, anticipating disruptions, automating operational responses, and supporting preventive maintenance. Computer vision tools are being used to detect suspicious behaviour, abandoned objects, and overcrowding; integration with MaaS and digital mobility passes is advancing to deliver more seamless services. Alongside these gains, participants stressed that AI introduces governance questions around purpose, proportionality, and transparency.

There was a consensus that the central question for transport authorities is not what AI can do, but what it should be used for. Five decision points were outlined for authorities considering AI deployment: (1) scoping, clarifying objectives and mandates and deciding whether to use AI at all; (2) risk assessment, ensuring outcomes are no worse than without AI and matching safeguards to risk; (3) technology selection, comparing AI and non-AI options and checking alignment with social, technical and legal expectations; (4) deployment oversight, integrating AI with human decision-making (assist, suggest, replace) while managing in-use risks; and (5) evaluation and reversibility, assessing expected versus actual performance and keeping open the option to roll back to human-centred systems if required.

Skills and organisational capacity were identified as enabling conditions. Participants emphasised the need for proportionate training and AI literacy across roles, so that staff can work with AI tools and critically evaluate their results. Not every automatable task should be automated: authorities should consider the long-term effects on professional judgement and institutional knowledge.

Examples from operators illustrated both progress and limitations. Predictive maintenance has been used for years, and more recent initiatives include real-time translation for passenger information, internal AI assistants for staff and computer vision to detect passengers remaining on trains at terminal stations. While these initiatives are advancing, practical constraints remain, such as raising digital awareness across the workforce, ensuring compliance with emerging AI rules, and continuing to improve data quality and governance.

From a European policy perspective, data and AI are seen as key tools for achieving the green and digital transitions. The Data Act and the development of a mobility data space aim to reduce fragmentation and enable the fair and secure sharing of data. Alongside the AI Act, the Commission is preparing measures to assist operators in transitioning from pilot schemes to everyday use, through training, testing facilities, and guidance that promote both technical adoption and responsible usage.

Beyond mass transit, beyond city limits: Governing the mix

The second part of the meeting addressed mobility outside dense urban areas. Participants noted that many residents remain 'stuck' in car dependency because alternative options are scarce or infrequent, particularly in rural and suburban areas, and at times of low demand, such as at night and during off-peak hours. Meeting these needs requires action at a regional level, as well as reliable, efficient and affordable services. There is no single solution, but rather a combination of services is required: mass transit where it is most efficient, alongside demand-responsive transport, shared taxis, carpooling, and other forms of shared mobility elsewhere.

Participants emphasised that this is not simply a matter of swapping vehicles. It involves service design and integration within the public transport offer, fare policy, sustainable funding models, the role of technology and data, procurement approaches, and monitoring and evaluation. These are governance tasks. Tactical governance includes engaging users in issues such as safety, reliability, ease of use and cost, shaping rules relating to fares, branding and satisfaction, aligning budgets across authorities and linking hub development with spatial planning. Operational governance covers the day-to-day delivery of travel information, reservation and payment systems, and planning.

Shared experiences pointed to practical pathways. The integration of carpooling into hub structures is advancing, with models in which users can access carpool trips using public transport passes, and in which drivers receive compensation per passenger. Such integration reduces uncertainty for users and treats carpooling as part of the public offering. However, participants cautioned that 'stand-alone' fixes can displace other services, and that coordination is needed to ensure an efficient mix across different geographies and times. Data-supported planning, potentially aided by AI, can help to match supply and demand and prevent unintended substitution.

Funding was discussed as a cross-cutting issue. Suggested options included employer contributions (as seen in regional models), engaging real-estate developers and better targeting of existing resources. For example, funds could be shifted from persistently underused fixed lines to integrated alternatives, while safeguarding core mass-transit capacity. Social instruments, such as providing younger residents with education on mobility costs and utilising national and EU funds, can prevent households from becoming burdened with high car ownership costs.

Conclusion

The meeting emphasised two complementary areas of governance. Firstly, when dealing with AI, authorities should have clear objectives, proportionate risk management, and organisational capacity. They should document their choices and performance while preserving reversibility. Secondly, achieving sustainable mobility beyond city limits requires an integrated service mix, regionally coordinated governance, and funding approaches that balance efficiency with accessibility. Moving forward, these discussions will inform POLIS's ongoing work on governance for emerging mobility systems.