Public transportation has always been about more than just moving people from point A to point B. It is the foundation of equitable, sustainable, and thriving communities. Yet commonly known challenges — like service reliability, fare structure and evasion, aging infrastructure, and dwindling funds — have plagued transit systems for decades. In the best case, those challenges can become the source of jokes in a local community; in the worst case, they become accepted as necessary evils.
In the age of AI, communities can and should challenge the status quo of these issues. AI is no longer a futuristic concept reserved for technology companies. It’s rapidly becoming a practical, essential tool for transit agencies to leverage to deliver safer, smarter, more rider-focused services. From optimizing routes and predicting demand surges to powering real-time multilingual chatbots to guide passengers, AI has the potential to redefine how riders interact with public transportation and how transit agencies manage their systems. Recent research from Deloitte found that 70% of transit agencies already plan to increase AI investments in the next 3 years.
Recognizing AI’s potential to transform public transit journeys and agencies, Cubic has partnered with AIDA Lab at Imperial College in London to track and work on emerging new technologies that are relevant to the wider transport domain. Drawing together inputs from our academic partnerships, product partner ecosystem, Cubic team members, our transit agency partners and the riders that use our systems, we’ve compiled a list of the top opportunities for AI to meaningfully improve the public transit experience — personalized trip help, reduced fare evasion, predictive maintenance, equity insights, fraud detection and schedule optimization
Personalized, Real-time Journey Assistance
With AI, transit agencies can provide riders with contextual, step-by-step support throughout their entire journey. AI-powered digital assistance — like chatbots, voice assistants, and in-app features — can easily guide riders through trip planning, transfers, delays, and payments. Thanks to advancements in generative and agentic AI this support can now go well beyond just answering a query or notifying you of a delay. AI can learn from your travel history, patterns and preferences to generate highly personalized, proactive suggestions to improve your entire journey.
Imagine, an accident has occurred on a major thoroughfare in your city. Lanes are blocked, traffic is backed up and buses into the city are delayed in the gridlock. By the time your alarm goes off in the morning, you learn about the accident and check for transit delays you could already be late for work.
Now imagine a scenario where AI helps plan your recommended journey before the day even starts. Should you wish, you could be woken earlier to ensure you make it to work on time, avoiding road closures or travel delays. Prefer to sleep in? You could proactively receive personalized recommendations for alternative routes that consider your preferences for modes, transfers, costs or crowding. AI powered tools could even help you book a seat or buy your fare.
Through partnerships with advertisers, your local transit agency could help encourage you to travel at a certain time, or on a certain route by providing you with discounts or deals, relieving pressure on the transport system while making your journey more comfortable. Your participation and input would influence the options you are provided. You can choose the information you’d like to share and how flexible you can be about your journey. reflecting the information, you choose to share and how flexible you are. Your participation and input would influence the options you are provided with, allowing you to make informed travel choices. On a larger scale this can benefit not only you, but the wider transport network and your city.
Personalized features can be especially useful for occasional riders, tourists, or multilingual users. AI customer service can drive improved communication and wayfinding for people with disabilities, access challenges, or local language limitations. The result: reduced uncertainty for commuters, more personalized travel experiences, greater access to services, and even mode shifts by incentivizing certain behaviors like off-peak travel.
Tackling Fare Evasion
Fare evasion is a costly, complex challenge for agencies. Globally, it accounts for billions in lost revenue — money that could otherwise be reinvested into services. While some fare evasion is obvious such as jumping a fare gate, other forms are much more subtle and harder to detect such as “short ticketing”.
AI can help transit agencies address these issues by detecting discrepancies between a passenger’s declared journey and their actual travel. Machine learning models can classify short ticketing behavior, highlight loopholes in fare types and policies, and pinpoint high-risk stations where agencies could benefit from increased staffing.
Real time edge-based AI computer vision processing is also being used to help detect physical fare evasion. Computer vision, implemented at gate lines or on buses, can help agencies detect and categorize many current forms of evasion including avoiding or delaying validator taps, using invalid tickets and exploiting back-door boarding.
The insights generated by these tools can enable agencies to identify trends, allocate resources efficiently and make data-driven decisions on how to optimize operations, enhance security and protect revenue. They can also help drive passenger compliance through greater real-time detection and reporting.
The application of these AI computer vision models can extend beyond fare evasion and could help address the rising frequency of safety incidents on public transit. While currently trained to detect fare evasion, these systems can also be trained to detect security events such as antisocial behavior.
Predictive Maintenance and Asset Management
For decades, transit maintenance has been reactive. Equipment gets fixed once it fails. But breakdowns in gates, ticket machines, or escalators don’t just cause inconvenience; they can directly reduce ridership and revenue.
AI improves the situation by enabling predictive maintenance. By combining computer vision, IoT sensors, and analytics, transit agencies can detect issues before they cause breakdowns. For example, AI can identify unusual vibration patterns in a fare gate motor or abnormal wear on an escalator belt. Replacement parts can be ordered before they are needed, and maintenance teams can focus on carrying out proactive maintenance rather than conducting multiple trips to diagnose the issue and then repair it. Predictive maintenance is key to avoiding service disruptions, reducing repair costs, and extending asset life.
Research conducted by McKinsey suggested predictive maintenance could reduce overall maintenance costs by 18-25% while cutting unplanned down time by up to 50%. Soon to be released research from Cubic and AIDA Lab echoes these benefits, suggesting predictive models could reduce the number of engineer on-site visits by up to 70%. For agencies under constant financial pressure, those efficiency gains are game-changing.
Smart Fare and Equity Insights
Fare systems are a powerful lever for ridership and equity, but agencies often lack the data to predict how changes will impact different communities. AI can analyze transaction data at scale to model the effects of different fare strategies — from fare capping to distance-based pricing or even dynamic pricing.
By simulating how different rider groups respond to fare changes, AI can help agencies design inclusive fare structures. This ensures that discounts or assistance reach vulnerable populations, that concessions are targeted effectively, and that fare policies encourage mode shifts instead of penalizing occasional riders.
The outcome is a more equitable fare system, improved access for underserved communities, and healthier revenue models for agencies.
Scheduling and Route Optimization
Transit demand is dynamic. Weather, major events, and even social trends can reshape ridership patterns overnight. Traditionally, agencies have relied on static timetables or historic averages — often leaving them under- or over-prepared.
With AI, agencies can analyze historical ridership, real-time conditions, and external factors like concerts or storms to forecast demand and adjust services accordingly. Machine learning can recommend adding bus frequency ahead of a sporting event, rerouting trains during adverse weather, or dynamically allocating vehicles in response to emerging demand. AI can also support agencies analyze real time vehicle capacity and help prioritize full vehicles over emptier ones so that more riders make it to their destination on time.
These optimizations reduce crowding, improve service reliability, cut emissions by reducing fuel waste, and help eliminate transit deserts. In some cities, AI-driven demand modeling is already informing on-demand bus pilots. In these programs, dynamically routed shuttles respond to real-time demand, reducing wait times and expanding coverage.
Fraud Detection
As more transit agencies shift from closed loop smartcards to open loop ticketing systems, fraud detection and management has become increasingly important. When a rider taps their credit card on a bus or a train, our fare payment systems need to make a quick decision on whether this transaction is valid or not. Cubic’s customer experience research showed fraud was the number one concern for riders unwilling to use their credit or debit card for fare payment.
Together, Cubic and Mastercard are exploring how AI algorithms and machine learning models can support agencies in conducting fraud monitoring and analysis. These models leverage past transit rider patterns to automatically detect and flag card usage and account balance anomalies. AI can learn ‘normal behavior’ from legitimate payment card usage including common:
- Transaction patterns: average fares, frequency of use, time of day, common routes
- Geographic data: unusual locations of card use
- Device data: the type of device frequently used for payment
- User login attempts: regular login times, locations, devices
When transactions deviate significantly from these patterns it can be flagged as suspicious, assigned a risk score and then escalated accordingly. Suspicious behavior could be events such as a single card being used multiple times within a very short period at distant locations; when a commuter who always uses the same device suddenly starts using a new one from an unfamiliar location; when multiple payment attempts are made with incorrect details or when a transaction volume occurs that doesn’t align with the riders typical fares or common journeys.
Unlike older, traditional rule-based AI systems, today’s AI models continue to learn and adapt to new fraud tactics and patterns, enhancing its ability to detect evolving threats. AI’s use in fraud detection will be critical in the “arms race” against networks of fraudsters who are frequently using AI too.
Challenges for Public Sector
Despite these clear opportunities for AI to benefit travelers and transit agencies, the public sector undoubtedly needs to overcome critical considerations to tap into the potential of AI.
- Implementation costs: Hardware, software, and integration require upfront investment.
- Data privacy and security: Safeguarding rider data is critical for compliance and public trust.
- Legacy integration: Older systems may require complex retrofits.
- Workforce impact: Staff retraining and adaptation programs are essential to avoid displacement.
- Ethics and regulation: Bias in AI decision-making remains a live concern, as do evolving regional regulations.
AI is not a silver bullet — but it does offer a powerful set of tools that can help agencies reimagine long-standing transit challenges. By piloting projects in fare evasion detection, predictive maintenance, and journey personalization today, agencies can build the foundation for smarter, more resilient transit systems tomorrow.
At Cubic, we’ve seen firsthand how AI can drive tangible improvements for agencies and riders alike. The future of transit isn’t just about moving people — it’s about making every journey safer, smarter, and more inclusive. Learn more by contacting our team.