Can AI solve the critical issues in the infrastructure industry?

Could AI be the cornerstone for resolving -finally- critical issues in the infrastructure industry? We have often heard, over the past 20 years, that the infrastructure sector is ready for a profound transformation of its value-chain in order to increase efficiency, effectiveness, and, in ultimate analysis, productivity. Indeed, it is well known that the general productivity in the construction industry has not shown any significant improvement in recent decades. The current conditions in the sector globally, however, lead us to think that this industry transformation may finally be approaching. Let’s consider the following: 1.  Global transformations and themes such as defense and energy transition are bringing unprecedented levels of funding to the sector. It is estimated that 139 trillion dollars of investment in sustainable infrastructure are necessary globally to achieve the objective of “Net Zero” by 2050, while an additional 94 trillion dollars must be invested by 2040 to close existing gaps and get aligned with future economic changes. 2.   At the same time, the sector is experiencing enormous uncertainty and instability. On the political level, beyond the obvious problems deriving from current wars and instabilities, more than half the world’s population has experienced a change of government this year, bringing additional uncertainty around the policies that will be implemented and the direction of future investments. This geopolitical situation is also having an impact on global trade and causing an upward trend in prices in the infrastructure value chains. 3.   Technology, and in particular the disruptive effect of the new Generative AI systems, is entering our daily lives at unprecedented speed, offering new potentials and solutions that seemed unthinkable only six months ago.

We therefore find ourselves in the context of an industry that historically has not shown a great amount of innovation, with large amounts of available capital, pressure to innovate deriving from the need to reduce risks, and new technological solutions that push us to improve the efficiency of our processes and production chains. We thus have all the necessary ingredients for a paradigm shift in the infrastructure sector. It is important at this point however to learn a lesson from recent history. In the past decade, for example, we have witnessed a race to invest in technology and data to transform our cities into smart cities, driven by new technological solutions and strong urbanization pressures.

However, the issue we are facing today, in many cases, is how to use these technologies and investments efficiently and effectively for the actual services provided to, and services utilized by the end users. Very often, in fact, these transformations have focused on systems and technologies but not on genuinely transformative solutions. Hence, while it is true that AI applications in infrastructure are growing at an exorbitant rate -and will continue to do so-, we must make sure that the resulting opportunities will not be a new race toward the mere implementation of innovative technologies in an otherwise static system.

Specifically, to gain understanding of the balance between “what is possible”, “what is real”, and “what ultimately brings value”, EY, in collaboration with FIDIC – International Federation of Consulting Engineers –, has conducted the study “How Artificial Intelligence Can Unlock a New Future for Infrastructure”. This global study focuses on fundamental aspects, including how AI is currently used in infrastructure, obstacles to its implementation, what needs to be done to allow broader use, and its potential benefits. From our understanding of what technology can offer (what is possible), what is actually achievable today (what is real), and the solutions that we see as transformative and value-adding (what ultimately brings value), we have distilled five areas of development that will facilitate a further convergence of systems and technologies into solutions with value: 1.  AI for “identifying the purpose” – Using data analysis and predictions of social, economic, and behavioural development can lead to increased efficiency of investment to meet the needs of the final users, both present and future; 2. AI for “end-to-end process planning” – Improving the efficiency of planning, financing, design, construction, and operation of infrastructure assets is clearly considered as one of the most evident area of integration and growth in the sector; 3. AI for “reworking the operating model” – Continually reassessing operating models and traditional supply models to meet constantly evolving infrastructural needs; 4. AI for “integrating work methods” – Cultivating a culture of collaboration and continuous learning, emphasizing interdisciplinary competencies and AI literacy at all levels of organization to improve results; 5. AI for “managing smart assets” – Ensuring that specific knowledge of current and future performance of assets is an integral part of decision-making processes by infrastructure operators.

These five guiding principles have been chosen for their broad reach and their ability to guide us toward the development of solutions with deeper efficiency throughout the industry. This approach however cannot ignore the requests and expectations of infrastructure operators and users seeking greater effectiveness and safety on the one hand, and a better use experience, on the other. If we look at transportation infrastructure for example, the needs of operators and users (travellers) can converge via the use of AI solutions along two general elements: •   Asset Management
Efficiency and resilience of infrastructure assets resulting in improved user safety; •    Traffic Management
Optimization of traffic management to reduce travel times and increase travel comfort.

In terms of asset management, predictive analyses based on machine learning based on different types of data (e.g., historical, environmental, sensor) can transform maintenance from planned and reactive to predictive, with positive outcomes for infrastructure efficiency and safety. As an example, intelligent systems have been tested on road infrastructure in the form of onboard surveillance vehicle technology, assessing the qualitative conditions of the road surface and identifying and classifying signage and its state of preservation. Efficiency and safety objectives can also be achieved by focusing interventions on maintenance and safety operations, with AI solutions guiding operators to identify the best procedure to implement and the best specific actions to undertake. In the realm of traffic management, AI-based systems enable smart planning, monitoring, and management of traffic and, in general, the use of infrastructure. For the rail system, AI algorithms make it possible to optimize timetable design, with the objective of maximizing capacity use and minimizing the impact of possible anomalies (delays, malfunctions). They also provide indications to operators regarding how to reprogram circulation in the event of a disruption of service. For the roads’ system, intelligent traffic systems use smart traffic lights to regulate traffic on the basis of vehicle flows detected by sensors, thus allowing for reduced congestion, and better distribution of vehicle flows. In airports, AI algorithms enable optimal planning and assignment of resources (e.g., check-in desks, gates, airplane parking stands, baggage belts).

An interesting development made possible by AI algorithms is the possible convergence of asset and traffic management to achieve overall optimization of the infrastructure. An optimized decision-making process could ultimately be guided by a value framework that accounts for economic factors (investments, maintenance and operating costs), operational factors (availability, route times, capacity), social factors (accessibility, safety, security, user satisfaction), and sustainability criteria (energy efficiency and environmental impact). Our privileged position as global partner of the largest and most innovative technological and infrastructural development programmes allows us to see what has effectively been implemented and what can be successfully implemented in infrastructure developments. This allows us to understand “what is real”, “what is possible”, and “what ultimately brings value”. Indeed, it is these broader considerations of value generated for the project, the economy, and society more in general that can truly lead to a convergence – through a process of natural selection – of AI solutions that will transform the entire infrastructure industry.

EY, formerly known as Ernst & Young, is a global network of professional management consulting, audit, tax, transaction and education services.