IIHS survey: 53% of Super Cruise, 42% of Autopilot, and 12% of ProPILOT Assist users say they were “comfortable treating their vehicles as fully self-driving”
Drivers using advanced driver assistance systems like Tesla (TSLA.O) Autopilot or General Motors (GM.N) …
Context & Ripple Effects
IIHS flagged the problem back in 2019, when it argued that marketing names like Autopilot lead drivers to overestimate what their cars can do. This new survey puts numbers on that thesis: a majority of Super Cruise users and more than four in ten Autopilot users say they are comfortable treating the vehicle as fully self-driving, while only 12% of ProPILOT Assist users say the same.
The spread across systems is the story. Super Cruise tops the list even though [[a:878484|NHTSA's 2016 review of the system highlighted its driver-facing camera precisely because GM built it to keep eyes on the road]] — suggesting monitoring hardware alone does not fix misperception. The findings also invert the public mood captured by [[a:929890|AAA's 2018 survey, where 73% of American drivers said they would be too afraid to ride in an autonomous vehicle]]: the people actually using these systems are far more trusting than the general public.
First-order effects
- Tesla, GM, and Nissan each face a perception gap specific to their own product: Autopilot and Super Cruise users massively overestimate capability while ProPILOT Assist users mostly do not, making naming and driver-engagement design the immediate variables under scrutiny.
- IIHS now has hard survey data tying over-trust to specific brand names, giving safety advocates and litigants a documented basis for challenging how these systems are marketed.
Second-order effects
- The absence of publicly available safety data from Tesla and other makers means neither regulators nor buyers can check whether this over-trust translates into crashes, pushing NHTSA toward demanding disclosure as the cheapest corrective.
- Driver-monitoring design becomes a competitive differentiator: GM's camera-based attention tracking did not prevent the highest over-trust score, so rivals may compete on stricter engagement locks rather than on automation features themselves.
Third-order effects
- If the pattern holds, the industry moves toward standardized labeling and mandatory driver-monitoring requirements for partial automation — a governance response to the widening gap between what assistance systems do and what their names imply.
- Over-trust among actual users, contrasted with public fear measured by AAA, points to a bifurcated adoption curve where experience breeds confidence faster than safety evidence accumulates — the structural risk ex-NHTSA adviser Missy Cummings later described as systemic over-trusting of these systems.
The trend: Driver-assistance branding is outrunning user understanding, and survey evidence like IIHS's is building the case for regulators to standardize how partial automation is named, disclosed, and monitored.