���ϳԹ���

"We live in a world where AI is accelerating risk in our lives at an ever-increasing pace," Solera's Sean Ritchie says. "Everyone feels they’re being left behind.” (John Sommers II for Transport Topics)

SAN DIEGO — Most fleets across America use in-cab, video-based safety systems, although driver acceptance has long been an issue for carriers.

Some of the latest products to come on the market were on display at American Trucking Associations’ 2025 Management Conference &Exhibition.

The latest weapon in the fight to keep America’s truck drivers safe — and carriers’ legal fees down — is real-time video footage and risks monitored by artificial intelligence.

Use of AI is one of the key themes of MCE 2025. But at least one longtime safety system observer and advocate believes video-based safety systems need human intelligence as much as AI.

“I think there’s nothing more important to fleets than safety,” Sean Ritchie , vice president of fleet solutions at risk management technology software specialist , told MCE attendees. “Ensuring operators come home safely to their families is paramount. The truth is, we live in a world where AI is accelerating risk in our lives at an ever-increasing pace. Everyone feels they’re being left behind.”

Image

AI-centric video-based safety systems use algorithms to decide whether something is risky. When a moment is identified as risky, the video is clipped and sent to the cloud. That is then made available to a carrier’s safety team.

However, Ritchie said, “There are serious deficiencies in just using an AI video-based safety system.”

There should be three areas of concern for executives: false positives, risk totality and false negatives.

False positives are when AI says there is a concern, but there isn’t.

“No matter what AI-based, video-based safety system you have, there will be false positives. Anyone who is telling you there aren’t is lying to you,” said Ritchie.

Fleets are operating at their leanest levels ever, which means that their safety specialists must maximize the use of their time.

“Every second they’re looking at a video that does not contain a risk, is time they’re not looking at a video where there is a risk,” Ritchie said.

Even the best AI-only video-based safety systems have 2%-10% false positives. False positives also distract drivers and create credibility problems for the video-based safety systems, said Ritchie.

Some 20% of a fleet’s drivers create 80% of the risk. “It is imperative that your teams focus on that 20%,” said Ritchie.

Risk totality involves overestimating how much of the risk occurring is being captured. Ritchie argues that 30% of risky incidents are not captured by video-based safety systems that only rely on artificial intelligence.

The biggest problem arises when litigation is threatened or initiated, with the possibility of a nuclear verdict.

Trucking insurance costs have been on the rise during a time when budgets are under their biggest strain in many years, with insurers trying to figure out how AI fits into the bigger picture.

“The trouble is that you are liable for all the data presented to you,” said Ritchie.

Carriers are liable for, must coach drivers on and document anything on a video on their servers that may or may not have been captured by the AI-only video-based safety system, he said.

“Your defense can’t be ‘Your honor, my AI didn’t flag this,’ ” he said.

There will be a point when technology will advance to the point where risk totality is not a factor, he said, but that is only likely in three to five years’ time.

As a result, fleets must implement safety solutions that have human intelligence — with monitoring and coaching — built into use of the video-based safety system, said Ritchie.

Carriers must implement coaching by humans rather than just AI avatar-based coaching. Drivers must know that the coach is someone who cares about them, their family and has done the job before, said Ritchie.

“If we give that up, we risk giving up all the advances we’ve made” with video-based safety, he said.

“I believe we’re in this perfect storm” of lean fleets, a litigious environment and AI cameras missing risky moments, he said, adding: “As an industry, we have to navigate through the perfect storm.”