Transportation’s Future Is in AI

Artificial intelligence, or AI, has been improving safety and increasing efficiencies in public transportation for years. What’s new about it now, and a key part of the Implementing AI for Public Transportation and Mobility Solutions session at APTA’s 2024 Legislative Conference, is the role transportation agencies must take in shaping guidelines for its use.

Moderator Nathaniel P. Ford Sr., CEO, Jacksonville Transportation Authority, cited the recent USDOT Complete Streets AI Initiative as an inspiration for the discussion, featuring MJ Maynard, APTA vice chair and CEO, Regional Transportation Commission of Southern Nevada; Laura Chace, president and CEO, ITS America and member, USDOT Transforming Transportation Advisory Committee; and Dan Katz, VP of Public Policy, Haden AI, and former USDOT chief of staff.

AI is a toolset of software and data-collecting hardware (cameras, sensors, etc.) that yield information for decision-makers. Chace said its “super-power,” is automating data analysis, taking the numbers and words from reports, sensors, historical data, etc. “to analyze them and make decisions in a much more efficient way than a human possibly could.”

Katz stressed that AI systems don’t actually make decisions, but instead capture the facts, as in a red-light camera trap, then deliver them to police to review and act on. The AI “is just aiding humans in doing what they do—and doing it more efficiently.”

Maynard highlighted several AI safety improvements. Showing a video of a driver who stopped on a busy freeway and exited his vehicle, she explained that sensors picked up the slowdown and alerted the traffic management team to call first responders, “and we could do that before 911 was called. We have seen such an improvement in how we’re managing and responding to incidents.”

Chace cited other examples: traffic monitoring and security cameras to identify cars, pedestrians, and, soon, weapons; trip planning and service analysis, to assess potential route updates; and systems planning and equipment maintenance.

Every AI tool is dependent on data, which requires security and thoughtful policies for its use, the panel agreed. Chace said “it’s well past time for a national data privacy law,” while Maynard and Katz urged agencies to take active part in data guideline development conversations, especially with lawmakers.

“Aggregated data is going to be critical for the future of so many agencies,” Katz concluded.