Can AI Save Our Infrastructure?

Infrastructure refers to the essential physical and organizational structures necessary for the functioning of a society or enterprise. It includes components such as roads, bridges, water systems, railways, airports, and communication networks. Beyond physical assets, infrastructure also represents the foundation that supports economic activities and daily life.

Infrastructure plays a crucial role in today’s world.? It enables trade, powers businesses, and connects workers to their jobs, contributing to economic development.? Infrastructure is paramount for access to basic services like education, healthcare, water, and energy. Put succinctly, it enhances the quality of life, and if implemented and managed well, it can create competitive advantages for towns, cities and countries and attract sought-after talent and businesses, resulting in job creation, contributing to the local economy and generation of valuable tax revenues. Finally, intelligently designed and maintained infrastructure enhances resilience and reduces vulnerability to climate related events – something that’s becoming of increasing importance.

The lifecycle of any kind of infrastructure can be broken into four key major phases: planning and design, building, operating, and maintaining.? With Artificial Intelligence (AI) tools and technologies becoming more accessible and ubiquitous, the area of infrastructure is very well positioned for increased efficiency at every stage of its life cycle.

At the planning and design stage, one can build highly complex models as ‘digital twins’ to simulate various project designs and its impact on key required outputs.? For example, in the world of transportation, one can model highways and bridges and simulate the behaviour of these systems on myriads of operating conditions such as traffic flows at different times of the day or when impacted by extreme weather conditions.? Digital twins can also model the effect of accidents on traffic flow both locally as well as the ensuing secondary impact on connected roads and highways.? The output of such models can then assist in lane configurations, traffic signal placements and timings, and potential introductions of flexible lanes and toll roads.

At the build stage, AI can be very beneficial in all aspects of operations and project management.? From managing complex supply chains to staffing and implementing complex workflows, AI enabled tools can significantly improve productivity and lower project execution risks.?? For example, AI based algorithms could track project deliverables in real time, pre-order parts for the next phase of construction, schedule road closures and plan detours.? Generative AI enabled applications can learn from operational patterns and assist construction crews and engineers on site with various insights and recommendations during the construction phase.

During the operating and maintaining phases of an infrastructure asset, operating costs can be optimized using predictive algorithms that can recommend preventive maintenance at the right time to thwart catastrophic system failures, evade excess costs, and prevent service interruptions and other inconveniences to end users.? In my PhD thesis, I created genetic algorithm embedded computer models to design systems that needed to operate close to hundred percent reliability levels.? These complex models recommended the exact number and timing of preventative maintenance actions to minimize system life cycle costs.? The cost to run these models were significantly higher just a few years ago making it hard to deploy for infrastructure use cases.? However, with computational costs coming down significantly and with the advent of new hardware such as ‘inference chips’ one can cost efficiently deploy these pre-trained models to make real time decisions based on operational data from ongoing projects.

Looking ahead to how this process can get started, it’s important to note that the mobilization of AI initiatives requires preparedness at the organizational level that starts with investment in the right talent and skills.? As a start, an interdisciplinary team that is skilled in identifying, sourcing and handling data from various departments in the organization is required.? The right kind of data is a critical component of AI models, as it provides the input necessary to train said models to make accurate predictions. Over time, organizations’ proprietary data becomes their competitive advantage, a phenomenon also referred to as ‘data dominance.’??

We need to prioritize the broader implementation of various AI initiatives in mission critical projects that can both generate returns on financial metrics as well as benefit society as a whole.? At the city level, with municipalities plagued with budget constraints and rising costs for the maintenance of city infrastructure, AI has the potential to generate significant savings in operating costs through the implementation of proactive maintenance and bespoke planning strategies.? For example, in municipal planning, data on demographics, population growth, housing demand, and transportation could be used to predict future needs. AI can also analyze extensive datasets to prioritize maintenance and development projects related to roads, infrastructure, and utilities.

We are in the early innings of the AI revolution and as we start developing more tactical plans for implementation, considerations will of course need to be made regarding ethics, bias, fairness, and overall striking a balance between privacy and transparency.?

It’s essential to recognize that when a paradigm shift in technology like this occurs, we need to seize the moment. Infrastructure is a perfect place to start this AI journey given the potential for improving the lives of millions of people globally. So, let’s embrace this transformative moment!?

The opinions expressed here belong solely to me and do not reflect the views of any organizations with which I am affiliated.