In the year since ChatGPT launched, the AI discourse has gone in every possible direction. Creative types, for example, worry about the end of the publishing industry: if a simple prompt can generate a 500-page manuscript, why pay authors? Meanwhile trucking leaders are wondering how it might affect our industry in the future. Naturally, STC has a few ideas.

Transportation, with its continued push for improved efficiency through technology, was already in the AI game. Route planning or other optimization software dates back at least 25 years, depending on your definition. The first programs typically relied on continued human input. They were often disconnected from other parts of the company and from external systems that could monitor and anticipate changing weather patterns and other variables that could upset the supply chain. Today’s systems have evolved to crunch data at a highly sophisticated and complex level, helping the industry manage data in new ways.

ChatGPT is a sub-set of the broader AI universe.  (The basic premise is that it’s a really smart chat box – the kind you see on the Help pages of websites – but seriously, it’s very, very smart.)  It performs its function by mimicking a human conversation. You ask it a question; it gives you an answer.  But whereas Google and other legacy search engines also give you answers, ChatGPT doesn’t just return a list of websites, it predicts the answers to the specific question you asked, by understanding the prompt and providing a string of words that it predicts will best answer your question, based on the data on which it’s been trained.

STC questions if AI could be trained to read police accident reports and associated data.  From here, it could create realistic animations and other models to validate or refute the information contained in the report.  Drivers, carriers, insurance providers and other stakeholders could enjoy a more comprehensive picture when making determinations on preventability, rates, and the potential for litigation. It could also help law enforcement officers write and/or summarize reports based on information already available (weather conditions, traffic patterns, etc.)  If the AI can incorporate considerably more data points than an officer would reasonably have access to (or the time to research them),

A third use might be initial adjudication of challenges Crash Preventability Determination Program, in which the AI could review submitted material and make an initial determination. In fact, until recently, FMCSA’s Active Research website listed a study in which they were to investigate this exact use-case. STC would contend, however, that human validation and approval is still a critical responsibility.

STC has witnessed the introduction and evolution of various forms of technology into the transportation sector over the years. Some have been lasting; others died quickly (RIM pagers, anyone?), although they often were the precursor to something better.

The exponential rate at which AI continues to develop gives the impression that it’s not going anywhere anytime soon. Harnessing it for the betterment of the industry (not to mention society and humanity) will be a challenge worth watching. Used correctly, AI can transform our industry. Although STC would contend that human validation and approval are still critical, AI might help us improve safety.

(Wondering if this article is AI-generated? It’s not, but how would you know?)