Now that the Infrastructure Investment and Jobs Act has been signed into law, billions of dollars will flow toward an array of technologies intended to upgrade the country's infrastructure, from automated construction machinery to AI to detect lithium for clean energy development. And though they are not spotlighted, automated and artificial intelligence technologies are sprinkled throughout the highly-debated 2,700-page document.
The law could have a profound effect on how people drive and how they work, as it fuels all sorts of automated technology. At this stage, exactly how much of the law's $1.2 trillion will be used to invest in systems that employ automated decision making or AI cannot be known.
However, even though AI isn't core to the law, today's technology landscape demands that AI play a role, said Anton Korinek, a Brookings Institution fellow and professor in the Department of Economics and the Darden School of Business at the University of Virginia who studies the implications of AI for labor markets.
"Nowadays you wouldn't be able to make a significant investment anywhere without touching on machine learning, AI and automated systems," he said. But overall, the legislation "is not a concerted push for AI," he said.
"We would have liked to see more money being directed towards mandating the adoption of new technologies for capital infrastructure owners," said Balaji Sreenivasan, CEO of Aurigo Software Technologies, a company that provides enterprise software products used in large infrastructure projects. "Billions of dollars were allocated towards the rebuilding and repair of roads, bridges, public transit, rail, airports and wastewater facilities. Effectively planning, designing and constructing these initiatives will indubitably require intelligent technologies at every step of the way."
Still, despite some disappointment that the law does not direct more funds toward AI specifically, there are plenty of ways it will spur investment for emerging and sometimes unproven and controversial AI technologies.
AI for construction work, manufacturing and climate
Though automated systems are mentioned many times in the language of the law, the terms "artificial intelligence" and "machine learning" make few appearances. They do, however, show up in relation to climate change and advanced energy manufacturing.
The law carves out funding that could have a direct effect on the day-to-day lives of the people building or repairing roads, bridges or other physical infrastructure. Promising to make construction work faster, safer and more accurate, it allocates $20 million each year from 2022 through 2026 to fund the use of digital systems on construction sites such as automated and connected machinery.
Cloud data and software vendors are already building the underpinnings of a manufacturing industry and transportation infrastructure fueled by data and automated technologies in the cloud. Autodesk's software is a ubiquitous foundation of cloud-based construction design. Google Cloud and even former Google AI alums are finding ways to serve the manufacturing sector with industry-specific tools that detect parts defects. And AI models built using IoT sensor data from companies like Samsara promise to make factories more efficient or traffic smoother or safer.
There's also a nod to "smart manufacturing" in relation to industrial energy efficiency in the law. It calls for the Department of Energy to include manufacturing tech that employs AI, network operation automation and monitoring of the energy used by buildings in its industrial research and training programs. And it allows states to use related funding to help provide high-performance computing resources to small and medium businesses.
The law even funds AI and machine learning technologies used to develop geologic models for detecting critical minerals (think lithium for energy storage). And it calls for a report assessing the use of AI and ML for "climate solutions."
Coming soon: new rules requiring automated driver assistance tech
The law appropriates more than $200 million each year from 2022 through 2026 to fund various vehicle safety provisions. It calls on the secretary of transportation, former Democratic presidential candidate Pete Buttigieg, to make rules requiring automated crash-avoidance tech in vehicles. That means in a few years all new passenger cars sold in the U.S. will come equipped with:
- Automatic emergency braking systems that stop a car if distances from a vehicle or object ahead close in too quickly
- Lane-assist systems that warn people when they drift out of their driving lanes and automatically steer them back
- Tech that monitors for drunk and impaired driving, which could include automated or AI-based systems.
These sorts of semi-autonomous vehicle technologies are already installed in some new cars on the road today. However, Nandita Sampath, a policy analyst with Consumer Reports focused on algorithmic bias and accountability issues, said many autonomous vehicle features are not reliable.
"It's very clear that many autonomous vehicle companies are struggling with very basic automation," she said. "We need some really strict testing standards before this stuff becomes a requirement."
In addition to the new rules, the law funds research of driver-monitoring systems, and testing of autonomous vehicles on public streets.
Cellular-to-everything sensors detecting things and people
While many of the provisions of the sprawling law do not mention particular forms of automated or data-driven technologies, it does specifically call for use of cellular vehicle-to-everything tech, which relies on sensors in vehicles and other objects to detect other vehicles, bicycles and people. The law says that states can apply funding to projects that deploy vehicle-to-infrastructure communications equipment, which is sold by companies like Ericsson and Qualcomm and promises to help reduce traffic congestion to limit carbon emissions and prevent autonomous vehicle accidents.
Cellular-to-everything is a component of the connected tech landscape many backers believe a 5G-enabled city infrastructure should include. But like other AI and automated systems funded by the new law, it spurs fear of even more dystopian surveillance tech than our phones already facilitate.
Not only does cellular-to-everything tech produce new forms of location-tracing data that could be used by law enforcement, said Liz O'Sullivan, CEO of AI compliance and governance platform Parity, "if all of these cars and devices are networked, it gives the device producers access to a unique amount of information about the ways we move about the world. This information could be very valuable to these mega-corps, giving them an unfair advantage in the proliferation of, for example, smart cities."
In addition to funding use of automated traffic enforcement systems by states for work or school zones, the law also funds a "SMART grant'' program – short for Strengthening Mobility and Revolutionizing Transportation. The program will fund projects related to the use of autonomous vehicles, sensor-based infrastructure and unmanned aircraft systems like drones for traffic monitoring and infrastructure inspection.