The promise of artificial intelligence in urban environments is vast: from optimising transport networks and reducing carbon emissions to enhancing public safety and improving citizen services. Yet the journey from aspiration to operational reality is paved with data—clean, structured, interoperable data that AI systems can trust and act upon. As cities around the world begin to explore the potential of AI, a critical question emerges: are our urban data ecosystems ready?
This question lies at the heart of a recent wave of smart city initiatives, with Sunderland in the United Kingdom emerging as a notable case study. Through a deliberate strategy of digital infrastructure investment, cross-sector collaboration, and a focus on low-carbon innovation, Sunderland is positioning itself as a model for how cities can lay the groundwork for AI before the technology itself takes centre stage.
Why Data Groundwork Matters
Artificial intelligence models are only as good as the data they consume. In a city context, data flows from countless sources: traffic sensors, streetlights, weather stations, social media feeds, energy meters, public transport ticketing systems, and citizen feedback platforms. Without careful curation, this data is often siloed, inconsistent, or incomplete. AI systems trained on such data risk producing biased, inaccurate, or unreliable outcomes—potentially undermining public trust and wasting public funds.
The concept of data groundwork refers to the systematic preparation of urban data assets for AI consumption. This includes standardising formats, ensuring metadata quality, establishing data governance frameworks, creating secure sharing mechanisms, and embedding principles of interoperability, inclusivity, and human oversight from the outset. As Cristina Bueti, an expert from the International Telecommunication Union (ITU), explains, cities must prioritise these elements now—before fragmented systems and vendor lock-in define the future of urban AI.
Bueti’s call to action resonates strongly in the context of digital twins. A digital twin is a virtual replica of a physical city, fed by real-time data streams and capable of simulation and predictive analysis. These twins are increasingly seen as the intelligent operating layer for urban infrastructure, enabling managers to test scenarios, optimise resources, and respond to emergencies. But without a solid data foundation, a digital twin is little more than a hollow shell.
Sunderland’s Smart City Journey
Sunderland, a city of around 275,000 people in North East England, has long been known for its industrial heritage—shipbuilding, coal mining, and automotive manufacturing. In recent years, however, the local authority has embarked on an ambitious transformation to reinvent the city as a hub for digital innovation and low-carbon growth. Central to this vision is the creation of a smart city ecosystem built on data and AI readiness.
The city has invested heavily in digital infrastructure, including a city-wide fibre network, Internet of Things (IoT) sensor deployments, and open data platforms. These assets are not ends in themselves but enablers for a range of AI-powered applications: smart traffic management, predictive maintenance of public assets, waste optimisation, and energy efficiency. By making data accessible and interoperable from the start, Sunderland aims to avoid the trap of vendor lock-in and create a competitive marketplace for AI solutions.
A key component of Sunderland’s strategy is the development of a digital twin platform. This twin integrates data from multiple sources—building management systems, transport networks, environmental sensors—to provide a holistic view of the city’s operations. City planners can simulate the impact of new developments, traffic flows, or climate events before committing resources. For example, the twin can model the effect of a new bus lane on congestion, air quality, and pedestrian movement, allowing more informed decision-making.
This work aligns with broader trends in the smart city sector. In Dublin, for instance, authorities are using digital twin projects to reduce traffic congestion, improve economic growth, and enhance community services. The Irish capital has deployed sensors on key corridors and integrated data from multiple agencies to create a living model of the city. Similarly, in cities like Barcelona and Singapore, digital twins are being used to manage everything from energy grids to disaster response.
Interoperability and Standards: The Critical Enablers
One of the most significant barriers to AI readiness in cities is the lack of interoperability between different systems and data formats. A traffic sensor from one vendor may output data in a proprietary schema, while a weather station from another uses a different standard. Without common data models and APIs, the cost of integrating these sources becomes prohibitive, and AI models trained on one dataset cannot easily be transferred to another city.
Cristina Bueti’s work at ITU focuses on developing international standards for smart cities, including those related to data interoperability. She argues that cities must adopt open standards and reference architectures to ensure that data can flow seamlessly across applications. This includes standardising data on air quality, energy consumption, mobility, and public safety. When such standards are in place, AI solutions can be scaled more easily, and cities can avoid being locked into a single vendor’s ecosystem.
Heinz von Eckartsberg, an architect and urban strategist at Woods Bagot, and Pablo Sepulveda from Impact Future, have taken this concept further by advocating for 'upstream resilience' in urban design. Their approach emphasises the need to design physical and digital infrastructure together, ensuring that data collection points are integrated into buildings and public spaces from the planning stage. By doing so, cities can capture high-quality data from the outset, rather than retrofitting sensors later. This upstream thinking reduces costs, improves data accuracy, and enables AI systems to operate with richer inputs.
Transport Networks: A Rich Data Environment
Urban transport is one of the most promising domains for AI application. Traffic congestion costs economies billions of dollars annually, not to mention the environmental and social costs. AI can help optimise signal timings, predict demand for public transit, and manage the integration of new mobility services such as ride-hailing and micromobility. However, the success of these applications depends on the quality and timeliness of transport data.
Sunderland has been active in this area, using data from its smart traffic system to improve journey times and reduce emissions. The city has deployed adaptive traffic signals that adjust in real time based on vehicle flows, and it is piloting AI-based algorithms to predict traffic patterns during major events or roadworks. These tools rely on historical data combined with live feeds from cameras, induction loops, and GPS probes. To make this work, the city has invested in a central data hub that normalises all inputs into a common format, enabling the AI application to access a unified view.
Similarly, Dublin’s transport authority has integrated data from buses, trains, trams, and private operators to offer real-time journey planning and demand forecasting. The city also uses AI to identify maintenance needs on its fleet, reducing downtime and improving reliability. These examples illustrate that the data groundwork for AI in transport involves not just technical integration but also institutional collaboration—breaking down silos between different agencies and private partners.
Smart Lighting: From LEDs to Data Hubs
Streetlights are one of the oldest and most widespread pieces of urban infrastructure. Once a simple means of illumination, they are now being transformed into smart nodes that host sensors, cameras, Wi-Fi access points, and even environmental monitors. This evolution is part of a broader trend toward 'cities thriving on lighting', as explored in a recent podcast series by SmartCitiesWorld and Paradox Engineering.
Sunderland is among the cities modernising its streetlight network, replacing older lamps with LEDs and adding control systems that allow remote dimming and fault detection. The next step is to layer on additional sensors—for air quality, sound levels, footfall, and parking occupancy—and connect them through a secure, interoperable network. These sensors generate a stream of data that can feed into the city’s digital twin and be used by AI algorithms to optimise maintenance, reduce energy consumption, and improve public safety.
The challenge, as highlighted in the podcast, is ensuring that these lighting networks are secure and interoperable. Many early smart city projects fell victim to proprietary solutions that left cities dependent on a single supplier for upgrades and data access. By adopting open standards and modular designs, Sunderland aims to retain flexibility and avoid such lock-in. This approach not only reduces long-term costs but also allows the city to integrate new AI applications as they emerge.
AI and Spatial Intelligence: The Citiverse
Beyond individual systems, a broader vision is emerging for the 'Citiverse'—a digital ecosystem that combines AI, spatial intelligence, and virtual worlds to deliver trusted, people-centred outcomes. Paul Wilson, an advocate for this concept, points to events like the UN Virtual Worlds Day as opportunities to shape how these technologies evolve. The Citiverse would allow citizens to interact with city services in immersive ways, from virtual town halls to personalised mobility planners, all underpinned by AI that understands context and preference.
However, the data groundwork for the Citiverse is even more demanding. It requires not only interoperability across systems but also the ability to handle 3D spatial data, real-time location information, and personal data with robust privacy protections. Cities like Sunderland are beginning to explore this frontier, recognising that the data architecture they build today will determine what is possible tomorrow.
Indoor Safety and Sensor Networks
AI’s impact extends beyond streets and transport to indoor environments—offices, hospitals, schools, and public buildings. Smart sensor networks can detect early signs of fire, gas leaks, or structural vulnerability, improving situational awareness and supporting healthier, more secure spaces. Sunderland is incorporating such systems into its public buildings, using data to optimise ventilation, lighting, and energy use while also enhancing safety.
These indoor systems must be integrated with the wider city data infrastructure to be effective. For example, an alert from a building’s fire sensors should automatically trigger adjustments to traffic signals, bus routes, and emergency response coordination. Achieving this level of integration requires the same data groundwork: standardised data formats, reliable communications networks, and clear governance protocols.
The Role of Special Reports and Knowledge Sharing
As more cities embark on these journeys, the need for knowledge exchange grows. Publications and reports that distil best practices—such as those analysing Dublin’s economic growth strategies or Sunderland’s smart city profile—serve as valuable resources. They help city leaders understand what works, what pitfalls to avoid, and how to build a business case for AI readiness.
Ultimately, preparing for AI is not about procuring the latest algorithms or hiring data scientists. It is about building a data ecosystem that is clean, curated, connected, and governed by principles of transparency and inclusivity. Sunderland’s experience shows that by focusing on the data groundwork first, cities can unlock the full potential of AI to improve efficiency, resilience, and sustainability for communities.
Source: Smart Cities World News