NHTSA Wants AEB on Every New Semi. Its Own Investigators Found the System Slamming Trucks to a Stop on Empty Highways.

Automatic emergency braking is genuinely good technology. It sees the stopped traffic your distracted eyes missed, and it hits the brakes before your foot finds the pedal. NHTSA wants it on every new heavy truck in America, and projects 155 lives saved per year once the fleet turns over.^[1] Lovely. Now consider the other file on the agency's desk: its own defect investigators have tied false AEB activations in heavy trucks to 24 crashes, 13 injuries, and one fatality, and the investigation is still open.^[2]

One hand writes the mandate; the other investigates the technology. Both hands belong to the same agency, and they are not talking to each other.

In June 2023, NHTSA and the Federal Motor Carrier Safety Administration proposed requiring AEB and electronic stability control on every new truck and bus over 10,000 pounds. Their math: about 60,000 rear-end crashes a year involve a heavy vehicle as the striking vehicle, and the mandate would prevent 19,118 of them, saving 155 lives and 8,814 injuries annually.^[1] The 2026 regulatory agenda says a revised proposal is coming imminently, with a final rule possible in 2027 or 2028 and Class 7 and 8 trucks complying about three years after that.^[3]

One month before that 2023 proposal, NHTSA's Office of Defects Investigation opened a probe into the same technology. Preliminary Evaluation PE23010 covered 2017-2022 Freightliner Cascadias and Western Star 5700s, roughly 250,000 trucks, after 18 complaints of the AEB system firing "without an actual roadway obstacle."^[4] In October 2025 the probe was upgraded to Engineering Analysis EA25006, the stage right before a recall decision. By then investigators had reviewed 347 reports: 24 crashes, 13 injuries, one fatality.^[2] Some events were the brief five-mph decelerations Daimler Trucks North America describes. Others were full service-brake applications that brought an 80,000-pound rig to a complete stop in a live travel lane.^[4]

So the mandate and the defect investigation are the same age. One of them will write the rules for every new semi on American roads; the other is still trying to decide whether the technology is a defect. The timeline collision is the story, but the math gap is worse.

All 155 projected lives saved come from rear-end crashes where the heavy truck is the striking vehicle. Phantom braking creates the opposite kind: the truck decelerates hard or stops dead for nothing, and becomes the struck vehicle, a forty-ton roadblock on a 65-mph highway.^[4] That crash category appears nowhere in the benefit analysis. EA25006 gives a crude floor for the induced risk: 24 crashes per 347 reviewed false-activation reports, about 7 percent. NHTSA's "155 lives" is a gross number wearing a net number's clothes.

Then there's the system-level embarrassment. ODI compared the two AEB families on those Daimler trucks and found the newer Detroit Assurance 2.0, 4.0, and 5.0 systems failing at similar rates, while the older, simpler Wabco OnGuard showed a significantly lower claim rate on comparable models in comparable conditions.^[2] Three generations of upgrades. Same failure rate. The fancier the system, the more it slams the brakes at shadows, which is what you'd expect from safety tech designed by marketing departments.

To be fair to the agency, its own 2023 proposal admits the problem: the two false-activation tests in the rule (driving over a steel trench plate; driving between two parked vehicles) "establish only a baseline for system functionality" and are "not comprehensive, nor sufficient to eliminate susceptibility to false activations in the myriad of circumstances in the real world."^[5] At least the confession is in the paperwork. And the counterargument deserves its full strength: 19,118 prevented crashes dwarf 24 induced ones, 155 saved lives dwarf one lost, and phantom braking touched roughly 0.06 percent of the investigated truck population per year. The mandate is almost certainly net-positive. OOIDA's opposition, it must be said, comes from a lobby that also fought electronic logging devices and speed limiters.^[3]

But "net-positive" is not a license to skip the accounting. NHTSA has never published an induced-harm estimate for heavy-truck AEB, so the headline number is gross, and gross numbers are for press releases, not engineering. Nobody is asking the agency to abandon the mandate. The ask is smaller and harder: finish the investigation first, publish the net math, and stop writing rules for a system your own engineers haven't finished judging.

What you can actually do: if you spec or buy fleet trucks, ask which AEB generation ships on the vehicle; ODI's data favors Wabco OnGuard over Detroit Assurance 2.0/4.0/5.0 on claim rates.^[2] If your truck phantom-brakes, file a complaint with NHTSA's Office of Defects Investigation at safercar.gov; those 347 reports are EA25006's data supply. And if you drive a 2017-2022 Freightliner Cascadia or Western Star 5700, know that your truck's safety system can, in documented cases, fully apply the brakes in a live lane with nothing in front of it.

What this article did not prove: the "one fatality" link is an investigator-reviewed association, not a confirmed causation finding, and EA25006 has not published final conclusions. The 7 percent crash-per-report rate is a floor from reviewed reports, not a population rate; most false activations are never reported. NHTSA's benefit estimates assume full fleet saturation with AEB and ESC and exclude induced crashes entirely, which is the gap identified here, not a number this article can fill.