Editor’s note: The following is a guest post from Claire Rutkowski, SVP and CIO Champion at Bentley Systems.
OpenAI’s release of ChatGPT was a watershed moment for AI. Visibility and access to the technology elevated its profile from being used behind the scenes to something that cannot be ignored.
However, its release has raised concerns from all corners, with people asking whether AI can — and should — be used on infrastructure projects or in the operations and maintenance of existing assets.
Architecture, engineering and construction (AEC) firms and asset operators are concerned about safety, risks and job security. Some have also concluded that AI has no place in a safety-driven, risk-averse industry like infrastructure.
While we certainly need to ensure that AI is used appropriately and ethically, the reality is the technology has been deployed for years.
A common use of AI in infrastructure is computer vision. This technology lets users identify and classify objects in images and video, such as when drones are used to take pictures of bridges rather than having people out in rappelling gear under risky weather conditions.
ML is then applied to images containing cracks, spalling, corrosion or obstructions to make maintenance recommendations. Layer in all the data gathered from IoT sensors to detect leaks, temperature fluctuations and other data, and it is easy to see how AI can deliver safer and cost-effective outcomes that improve service reliability and asset performance.
Workforce impact
AI can help improve safety and reduce risk. But will AI also take away jobs? Goldman Sachs Group certainly thinks so. The firm recently released a study predicting that generative AI tools, such as ChatGPT, could make 300 million full-time positions redundant globally, with the United States, Northern Europe and the United Kingdom more impacted than other countries.
Agriculture, mining, and manufacturing jobs are predicted to be the least impacted, while fields using programming or writing skills such as IT, administration, the legal field, and marketing are likely to see major disruption.
However, the infrastructure sector is facing an acute talent shortage. While specific numbers are extremely hard to come by, workforce limitations are routinely cited as the biggest impediment to project delivery over the next three years. Leveraging tools that help workers tackle lower-value activities is critical so that they can focus on the higher-value ones.
If AI can help the industry by augmenting the productivity of designers, engineers and operators, it can be the technology that helps us meet the world’s growing infrastructure demands at a time when there are more projects than professionals to get the work done.
As technology advances, there will be more applications in all aspects of the infrastructure lifecycle. Instead of having teams of people responding to requests for proposals, much of the work could be done by AI.
Firms are already working on training large language models with all their previous responses to RFPs so AI can generate drafts, as a starting point for more expert editing and finalization. AI can be used in procurement to automate routine tasks, conduct spend analyses and help spot opportunities for more strategic and agile sourcing across the supply chain.
The infrastructure sector already uses forms of AI in generative components, generative conceptual design and traffic and pedestrian simulations. Wouldn’t it be great if AI could make suggestions on how to reduce a project or asset’s carbon footprint, recommend design changes to improve sustainability or identify ways to reduce risk?
AI can also be applied across multiple assets (or systems of assets) to help analyze complex scenarios and make recommendations on transportation routing, schedules, and traffic signaling. It could similarly be used to analyze historic data on energy usage to predict demand and control systems, ensuring energy availability and reducing energy waste.
Given the infrastructure sector’s acute talent shortages and the need for safety and risk reduction, AI has had a place — and will continue to have one.
However, there will be risks as AI technology advances. Data usage or sharing must be transparent and legal. Designs must comply with code. Data models must be verified, and responses must be constantly checked for bias.
Firms and organizations working across the infrastructure sector cannot blindly follow the technology without questioning whether the answers provided are technically accurate and correct.
AI has the power to become a force multiplier that will increase the efficiency of our workforce so that we can meet the infrastructure demands of the world today – and tomorrow.
