Robotaxis and self-driving technology are expanding across U.S. cities, but safety researchers say we won't have enough data to confidently compare them to human drivers for several more years. The Insurance Institute for Highway Safety (IIHS), an Arlington, Virginia-based research organization, released findings showing that while advanced driver-assistance systems are helping prevent crashes, fundamental questions remain about how safe autonomous vehicles truly are.

What Safety Features Are Actually Working?

The most effective safety technology currently available is automatic emergency braking, which detects slowing traffic or hazards ahead and automatically slows a vehicle to avoid a crash or reduce its severity. Beyond that, newer partial driving automation systems, which allow vehicles to handle lane tracking alongside acceleration and braking simultaneously, are also showing promise in real-world conditions.

"The other area, when we start talking about partial driving automation systems, where the vehicle may indeed be able to do lane tracking as well as acceleration and braking at the same time, those systems can also have a positive safety benefit," said David Harkey, president of the Insurance Institute for Highway Safety.

David Harkey, President, Insurance Institute for Highway Safety

However, researchers worry that these improvements could create a false sense of security. As vehicles become more capable, drivers may become overconfident in the technology and engage in riskier behaviors like distracted driving.

Why Don't We Have Enough Safety Data Yet?

The core challenge is simple: autonomous vehicles have logged only a fraction of the miles driven by human motorists. The National Highway Traffic Safety Administration (NHTSA) requires companies operating autonomous vehicles and robotaxi services to report certain crash data, but this information is still insufficient to draw broad conclusions about safety comparisons.

"We're going to need a lot more data over a more extended period of time to really assess whether those systems are safe or not safe, and how they compare to human drivers," said Harkey.

David Harkey, President, Insurance Institute for Highway Safety

Think of it this way: if human drivers have collectively driven billions of miles, autonomous vehicles have driven millions. That gap makes statistical comparison nearly impossible. Researchers need to see how robotaxis perform across diverse weather conditions, traffic patterns, and edge cases that occur rarely but matter for safety assessment.

How to Approach Self-Driving Technology Responsibly

• Treat Automation as a Tool: View advanced vehicle technology as a safety aid that improves crash prevention, not as a replacement for attentive driving. Drivers should remain engaged and ready to take control at any moment.
• Monitor Regulatory Updates: Stay informed about NHTSA reporting requirements and safety standards for autonomous vehicles in your region, as these frameworks continue to evolve as more data becomes available.
• Understand Technology Limitations: Recognize that even promising safety features like automatic emergency braking have limitations in certain conditions, and no system is foolproof in all scenarios.

What Timeline Should We Expect for Safety Conclusions?

Harkey emphasized that gathering enough data to make definitive safety determinations could take several more years. Public trust in robotaxis and autonomous vehicles will ultimately depend on proving the technology is as safe as, or safer than, human drivers. Until that evidence is conclusive, regulators and researchers will continue collecting real-world performance data from robotaxi services expanding in cities across the country.

For now, the message from safety experts is clear: the technology is moving in a promising direction, but patience and rigorous data collection are essential before we can confidently say autonomous vehicles have solved the safety puzzle.