Unlocking the potential of artificial intelligence for America’s infrastructure

 Since celebrating Infrastructure Week in May, the conversation around infrastructure development has taken on a new dimension—one that goes beyond bridges and highways: harnessing the transformative power of artificial intelligence (AI) to propel our infrastructure into the future.

AI serves as a catalyst for productivity in a sector often burdened by resource constraints and historical backlogs, providing a pathway to enhanced efficiency. While AI-enabled technologies have been in use for almost 30 years in the infrastructure arena, the increasing availability of data provides a profound opportunity for innovation and progress.

A prime example is roadway maintenance. Regular maintenance saves lives and extends the useful life of transportation infrastructure. However, before any maintenance can occur, a department of transportation must be aware that a problem exists, which requires costly and time-consuming road surveys.

That’s where data and AI can help. For example, Hawaii DOT used crowdsourced imagery plus machine learning models to automatically survey its road network for cracks, debris, damaged guardrails and more, detecting 930 problems per week across the state. New York City DOT applied a similar approach to automatically inspect all 1,650 crosswalks in the city and analyze the condition of paint lines within each crosswalk.

These same technologies, applied nationwide, were used to create a map of public roads in the United States, complete with paint retroreflectivity scores for roads within state capitals. The map can assist state and local departments of transportation as they prepare to meet new Federal Highway Administration (FHWA) minimum levels of retroreflectivity for pavement markings, set to take effect in 2026.

In these cases, AI acts as a valuable assistant, but not replacement, for infrastructure professionals, augmenting their capabilities rather than supplanting them. It can improve existing work, like accelerating the speed and efficiency of roadway inspections, and it also frees up valuable time and resources by automating tedious tasks and streamlining decision-making processes, allowing professionals to expand their productivity and focus on higher-value work.

Consider, for instance, the potential of AI in exploring alternative designs to meet evolving infrastructure needs. By generating and evaluating multiple design scenarios, AI empowers professionals to push the boundaries of creativity and innovation, leading to more resilient and sustainable infrastructure.

Moreover, AI holds promise in improving forecasting accuracy, a crucial aspect of infrastructure project management. By leveraging predictive capabilities, AI aids in anticipating costs, effort required and project timelines, mitigating the risk of budget overruns, delays and rework—a welcome relief in an industry often plagued by uncertainty.

Beyond project-level benefits, AI has the potential to address broader societal challenges, from sustainable development to climate resilience. By harnessing AI capabilities to model and simulate flood risk, materials use and more, those building the communities of tomorrow can navigate complex issues with greater agility and foresight, laying the foundation for a more resilient and sustainable future.

However, the integration of AI into infrastructure development is not without its challenges. Data accessibility and context emerge as critical prerequisites for AI deployment. Infrastructure data, often fragmented and complex, require standardization to unleash their full potential in AI applications. Domain-specific context is essential to ensure the relevance and accuracy of AI models in infrastructure settings.

Fortunately, strides are being made to address these challenges. Through technology concepts like infrastructure digital twins—digital versions of physical assets—infrastructure data are transformed into a standardized, usable format, primed for AI applications. Additionally, domain-specific engineering principles provide the necessary context for AI models, ensuring their effectiveness in real-world infrastructure scenarios.

As we continue towards a technologically advanced infrastructure landscape, it’s crucial for policymakers to consider how they can incentivize and integrate technology into our vast infrastructure projects. With the historic funding from the Infrastructure Investment and Jobs Act (IIJA), there’s a unique opportunity to do more with less and ensure the longevity of our infrastructure. In shaping the future of American infrastructure, AI will no longer be a scary buzzword, but a driving force behind our infrastructure’s efficiency, resilience, and sustainability.

So, let us embrace the transformative potential of artificial intelligence in shaping the future of our infrastructure. By harnessing the power of AI, we can unlock new possibilities, drive innovation, and build a more resilient and sustainable infrastructure landscape for generations to come. It is time to leverage AI as our ally in pursuit of advancing infrastructure.

Julien Moutte is chief technology officer of Bentley Systems and is the principal architect of the company’s technology directions. He has more than 20 years of technology leadership experience in startups, scaleups and large organizations. Prior to joining Bentley as vice president of technology in 2021, Moutte served as head of technology for SAP Marketing Cloud and as a member of the office of the chief technology officer with SAP Customer Experience. He also served as chief technology officer of Scytl, a platform for online voting; and Fluendo, the free software multimedia experts, which he co-founded in 2004 in Barcelona, Spain. Moutte holds a degree in computer science from Université Claude Bernard in Lyon, France.