The Risky Driving Your AI Dash Cam Never Flagged Can Still Feed a Nuclear Verdict
At the ATA conference this week a defense attorney and a camera vendor warned that AI cameras miss risky driving, the footage still exists, and juries will not accept the AI as an excuse. Here is what that means for a small fleet and its liability coverage.
What a trucking defense attorney told fleets this week
On November 13, 2026, FreightWaves reported from the American Trucking Associations conference in Charlotte on a panel every small fleet running cameras should hear about. Doug Marcello, a trucking defense attorney with Saxton and Stump, said the biggest verdicts rarely turn on the crash itself. Plaintiff lawyers look for proof the company had a pattern of ignoring risk, and he said AI dash cameras are now handing them that proof.
The problem is not the camera. It is the risky driving the camera recorded but the software never flagged, so nobody ever coached the driver. His advice was blunt. Treat every piece of footage you hold as something the other side will eventually see.

False positives get fixed, false negatives get ignored
Marcello shared the stage with Sean Ritchie of Solera Fleet Solutions, a vendor that sells video systems with a human review layer, which is worth keeping in mind. Ritchie asked the room how many fleets knew their camera's false negative rate. Every hand went down.
A false positive is a flag for risk that is not really on the video. A false negative is risky behavior on the video that the AI never labeled. According to FreightWaves, Solera's review of AI triggered clips found up to 30 percent carried additional risk the software missed, drowsy driving most of all. Solera has not published its method, so treat that as a vendor estimate.
Why does it happen? Fleets complained that early cameras flagged too much, so vendors dialed the sensitivity down. Fewer nuisance alerts also means more real events slip through unlabeled. No federal rule covers how footage gets reviewed, and Marcello said that with no government standard, juries will write one, a verdict at a time.
How a mock jury treated a 35 truck carrier
Marcello described a mock trial he ran with a Philadelphia focus group. The made up carrier had 35 trucks and one safety director. A driver came over a hill, plowed into stopped traffic and killed two people. The week before, that driver had several speeding and hard braking events nobody had reviewed. The mock jury was completely unsympathetic.
That is the reptile theory, which Marcello described as a strategy built on making the jury afraid of a company that does not fix what it knows about. Unreviewed footage of a drowsy or speeding driver fits that story even when it had nothing to do with the crash. The panel came days after a Dallas County judge upheld a $604 million verdict against C.H. Robinson, a case we covered in our post on what happens when a verdict blows past policy limits.
Thirty days of quiet footage next to six months of flagged clips
Ritchie said most providers keep video with no identified risk for about 30 days and anything with a flag for six months or longer. So a clip flagged for a seat belt violation sits in storage for half a year with the unlabeled drowsy driving in the same clip, and if a lawsuit follows, all of it is discoverable.

Some fleets respond by never analyzing their footage, figuring what they never looked at cannot hurt them. Marcello called that the discovery fallacy. The data gets discovered either way.
Why this lands on your truck insurance
When a wreck turns into a lawsuit, your commercial auto liability policy answers the claim, and only up to your limit. A nuclear verdict built on a pattern of unreviewed footage does not stop at $1 million. Everything above the limit is yours.
It cuts the other way too. A camera program with documented review and coaching is one of the few things a small fleet can show an underwriter that moves the conversation. We wrote earlier this year about the class action over AI dash cam flags used as evidence, and the lesson is the same. The camera is a records system. Records help when they show you acted and hurt when they show you did not.
FreightWaves noted neither panelist pointed to a court ruling that turned on an event the AI missed, so this is a warning, not a settled rule. Ritchie also said dropping cameras is no answer.
What a one truck or five truck fleet can do this month
Ask your camera vendor one question. Does video go straight into your coaching queue, or does a human check it for missed risk first? Ritchie described three setups. Some vendors send everything straight to the fleet, some use reviewers only to strip false positives, and some have reviewers catch both before the fleet sees it. If you are on the first type, you are the reviewer, so set time each week to spot check clips the AI passed as clean.
Then document the coaching. A short note that you reviewed a clip and talked to the driver is the evidence Marcello says juries want to see. And make sure your liability limits reflect the world those juries live in. Get a truck insurance quote in under a minute at our quote page, or call or text 423-264-4255 and talk with someone who works with owner-operators and small fleets every day.
Common questions
What is a dash cam false negative?
It is risky driving that the camera recorded but the AI never labeled, so it never reached anyone for coaching. A false positive is the opposite, a flag for risk that is not really on the video. At the ATA conference this week, Solera Fleet Solutions said its review of AI flagged clips found up to 30 percent held other risk the software missed, with drowsy driving the most common miss.
Can footage my camera never flagged be used against me in a lawsuit?
Yes. Defense attorney Doug Marcello told the conference that any footage a fleet holds is discoverable, flagged or not, and that skipping review to avoid discovery does not work. Providers typically keep unflagged video about 30 days and flagged clips six months or longer, so a clip with one minor flag can preserve unflagged risky driving alongside it.
Does a dash cam program lower my truck insurance cost?
A camera by itself does not. A camera program with documented review and driver coaching is something an underwriter can actually weigh, and it is also the record that helps defend you after a crash. Call or text 423-264-4255 and we can talk through how your current program fits your liability limits.
How do I get a truck insurance quote for a small fleet?
Use the quote form on our site or call or text 423-264-4255. You can get a truck insurance quote for liability, cargo and physical damage in one conversation, whether you run one truck or a handful.
Ready for a better rate?
We shop A-rated carriers against each other to find your lowest rate, fast. Under a minute to start, and no obligation.
Prefer to talk it through? Call or text (423) 264-4255 and a licensed agent will walk you through your options.