Editor’s note: This is Part 2 of a three-part series from TechUnited:NJ on autonomous vehicles in New Jersey. This piece looks at safety; Part 1 covered the economic opportunity and Part 3 covers mobility.
My mom may never feel comfortable in a driverless car. I get it.
If you’ve spent your entire life believing that the person holding the steering wheel is what keeps you safe, taking that person away feels like adding risk.
But New Jersey has a particular reason to look past that instinct.
A national highway report released this year ranks New Jersey 50th out of 50 states for urbanized-area congestion. New Jersey commuters lose about 85 hours a year to traffic. And last year, 582 people were killed on our roads, including 175 pedestrians.
Few states have more to gain from making transportation safer and smarter.
That’s why a bill now moving through Trenton to launch a driverless car pilot matters. It gets a lot right. It also has flaws that could quietly undercut the whole thing.
The question shouldn’t just be whether driverless cars sound scary. It should be whether they are safer than what we have today — based on real data, not fear. For more than a century, there was no alternative to a human behind the wheel. We accepted distraction, fatigue, drinking and bad decisions as part of driving. Now we have enough real-world experience to compare humans with cars that drive themselves.
Swiss Re, one of the world’s largest reinsurers, compared more than 25 million miles of fully autonomous Waymo driving with its database of human driving. Across all those miles, there were just two bodily injury claims. Statistically, human drivers covering the same distance would have produced 26. Overall, the autonomous vehicles had 92% fewer bodily injury claims.
Then the Insurance Institute for Highway Safety independently studied about 50 million driverless miles. It found 68% fewer police-reportable crashes and 81% fewer injury crashes per mile than human drivers in the same cities and years.
While encouraging, it doesn’t prove every driverless car is safe or the technology is perfect. That’s exactly why New Jersey should set the safety standard, not pick the technology.
I don’t particularly care whether the winner uses lidar, cameras, infrared sensors or something that has yet to be invented. I care about the safety record.
Government should answer this simple question: How safe does a vehicle have to be before we let it drive without a human?
The bill moving through Trenton gets much of the structure right. It creates a three-year pilot. During the supervised phase, a trained driver must sit in the driver’s seat, monitor the vehicle and be able to take immediate control. Operators need at least $5 million in insurance, and a vehicle must complete at least 50,000 miles on New Jersey public roads before operating without that driver. NJDOT must publish qualifying crash data and can suspend unsafe operators.
But 50,000 miles is a mileage standard, not a safety standard.
At the human crash rate used in the IIHS comparison, a police-reportable crash happens roughly once every 250,000 miles. Fifty thousand clean miles can simply be good luck. The real question isn’t how far the car drove. It’s how safely it drove.
Two changes would make the pilot stronger: replace the 50,000-mile trigger with a measurable crash-rate benchmark using the public data the bill already requires and allow supervised vehicles to test in the challenging zones the bill currently excludes.
And yes, these vehicles need to prove themselves here. Newark isn’t Phoenix. The Turnpike isn’t San Francisco. We have our own weather, roads, traffic, construction and pedestrians. But no one wants to be a guinea pig, and New Jersey doesn’t have to be one.
We’re not starting at mile zero. Waymo alone had logged more than 220 million fully autonomous miles by the end of March 2026. Test the technology here. Make companies prove they can handle New Jersey. But recognize that we’re testing technology with an extensive track record proven elsewhere. The current ban on operating pilot vehicles in school zones, construction zones and areas with high pedestrian collision rates deserves another look.
I live in Hoboken. What I want to know is how a driverless car handles a crowded crosswalk, a construction worker directing traffic or a kid stepping toward the street.
Those are the hard situations, and they are where safety matters most.
During that first phase, a trained driver is physically in the seat and able to take immediate control. With that safeguard, it’s precisely when we should learn how the system handles the hardest streets.
There have been failures. Waymo vehicles repeatedly passed stopped school buses in Austin, triggering federal scrutiny and a software recall.
But let’s make sure we put that in context: Austin’s school district issued more than 7,000 violations that year for vehicles illegally passing stopped buses. Just 24 involved driverless vehicles — a stark reminder of what we’re measuring against. Not perfection. Us.
The question isn’t whether a driverless car ever makes a mistake. It’s whether, measured fairly over enough miles, it makes fewer mistakes that hurt people. Plus there’s a difference no driver’s ed class can match: one software fix can change how an entire fleet behaves.
There’s something bigger at stake too. New Jersey continues to be an underdog in the innovation economy. In 2025, New York captured nearly 90% of the venture capital invested across New York, New Jersey and Connecticut combined. Yet we have the talent, universities, scientists, engineers, entrepreneurs and history. Bell Labs invented the transistor here.
What we need more of is a culture of possibility — a state that starts by asking, “How could this work?” before listing all the reasons it can’t. That doesn’t mean lowering the bar. It means setting a high bar and giving innovators a path to clear it. Set the standard. Test it here. Publish the data. Make the technology earn our trust.
My mom may still hesitate. The numbers will have to earn her trust too. But the alternative is staying 50 out of 50 on a list where we deserve better.
Let’s ride.
Aaron Price is the CEO of TechUnited:NJ.
