Doug Marcello
There is a painful irony embedded in every nuclear verdict involving negligent training.
The company trained. They held the sessions. They logged the hours. They filed the paperwork. They checked every box their compliance department put in front of them.
And the plaintiff’s attorney stood in front of the jury and said: “Ladies and gentlemen, they knew this driver had problems. They knew. And their answer was a generic video.”
That’s the irony. The training they did became the evidence against them. Not the absence of training—its inadequacy.
This is the negligent training problem, and it is the softest leg of the plaintiff’s triad. It’s also the one most fleets can fix fastest—if they’re willing to admit that what they’ve been doing doesn’t work.
The Forgetting Curve Is Not a Theory
I recently interviewed Dr. Gina Anderson, a doctor of education and founder of Luma Brighter Learning, about what the trucking industry gets fundamentally wrong about training.
She didn’t sugarcoat it.
Within the first hour of hearing content, most people forget nearly half of it. That’s the forgetting curve—established neuroscience, backed by decades of research. And the industry’s standard training model is designed to maximize it.
Three days of passive onboarding. Generic content. Everyone in the same room getting the same information regardless of experience level. Then radio silence until the annual refresher—or until somebody has an incident and gets a corrective video.
“I told them that. They should remember it.” Dr. Anderson hears this constantly from safety directors. But that’s not how the brain works. It’s not even close to how the brain works.
The Experienced Driver Problem
Here’s something that should give every fleet executive pause: the drivers with the most experience are sometimes the ones who need the most training.
Dr. Anderson explains this through neuroscience. The brain builds neural scaffolds through association. When information is tied to an emotional experience or a long-held memory, those connections are powerful. They’re also stubborn. If a veteran driver learned a technique 20 years ago that’s now outdated, that old association doesn’t just fade—it actively resists new information.
She compares it to eyewitness misidentification. People in police lineups sometimes identify the wrong person with absolute certainty because their memory is built on association, not accuracy. The same mechanism is at work when a veteran driver insists they’re doing something correctly because “that’s how I’ve always done it.”
Onboarding a new driver and re-training a 25-year veteran require fundamentally different approaches. One-size-fits-all doesn’t account for this. It can’t.
Five Principles of Training That Actually Works
Dr. Anderson’s framework isn’t theoretical. It’s grounded in learning science and already being implemented by forward-thinking fleets. The principles:
- Individualization. Assess prior knowledge and behavioral data before training begins. AI can build tailored learning paths based on performance data, driving history, and behavioral indexes. The technology to do this exists today. The barrier is adoption, not capability.
- Continuous delivery. The brain needs seven-plus exposures to build reliable neural connections. Effective training delivers content in spaced intervals across varied formats—not in a single dump. Timing matters too: some people are most receptive early morning, others late at night. Delivering education when the learner is most engaged dramatically improves retention.
- Proactive intervention. The industry’s default is reactive: incident occurs, corrective video sent. But sending a speeding video to a driver who already knows not to speed doesn’t change behavior. Proactive training identifies patterns in performance data and addresses them before they become incidents—and before they become exhibits in litigation.
- Authenticity. This is Dr. Anderson’s doctoral research area. Authentic learning targets the individual’s specific, current context. A generic parking lot video is forgettable. A just-in-time simulation of the actual parking lot the driver is heading to—informed by their readiness to change, their motivational profile, and the specific hazards of that location—that’s an experience the brain retains.
- Connection over content. True micro-learning engages all five interaction modes. The most powerful—and most neglected—is learner-to-self: mindfulness and awareness. Where is the driver’s mind when they’re driving? Are they present, or are they thinking about being away from their family? This is the level of engagement that actually changes behavior. Six-minute videos don’t reach it.
The Courtroom Test
Everything above matters for one simple reason beyond safety: it’s what the courtroom demands.
Plaintiff attorneys attacking negligent training don’t need to prove you didn’t train. They need to prove you didn’t train effectively. They need to show the jury that your training was generic when it should have been individualized. That it was a one-time event when it should have been continuous. That it was reactive when the data showed you should have been proactive.
Check-the-box training gives them everything they need.
Data-driven, AI-founded training takes it away.
When your training platform generates objective evidence—the individualized learning path, the performance data that informed it, the spaced reinforcement schedule, the proactive interventions, the measurable outcomes—you’re not defending with a binder of sign-in sheets. You’re defending with data.
That’s the shift from subjective expert opinion to objective evidence. That’s AI defense.
One of Dr. Anderson’s clients lived this. They switched from check-the-box to science-based, authentic, individualized training. When they went to court, they didn’t say “we trained them.” They said “let us show you what we did.” A multi-million-dollar exposure resolved for $250,000.
The Denuclearization Thesis
This is the first installment in the Denuclearization series. The thesis is straightforward:
Nuclear verdicts in trucking litigation are built on the plaintiff’s triad: negligent hiring, negligent training, and negligent supervision. Each leg is an attack vector. Each leg can be reinforced or dismantled with objective, AI-driven evidence.
Training is where most fleets are most vulnerable—and where the fix is most immediately available.
The learning science is established. The AI technology is operational. Platforms that can deliver individualized, continuous, proactive, authentic training—and generate defensible evidence of it—exist right now.
The question for every fleet executive, safety director, and risk manager is simple: Are you training to check a box, or are you training to produce results?
One approach feeds the nuclear verdict machine. The other dismantles it.
That’s denuclearization.
See the full interview with Dr. Anderson at TransportCenter YouTube or the podcast at TransportCenter Podcast
