Sources: Tesla engineers discussed adding eye tracking and better sensors to Autopilot to help alert drivers, were rejected over tech's ineffectiveness and cost
Tim Higgins / Wall Street Journal :
Context & Ripple Effects
This report extends the picture painted by earlier sources detailing turmoil inside Tesla's engineering ranks after Musk's fully autonomous Autopilot announcement: engineers flagged driver-alerting gaps early, proposing eye tracking and better sensors, and were overruled on effectiveness and cost grounds. That decision is the through-line connecting today's story to later scrutiny.
The rejected proposals map directly onto the criticisms that followed — experts and lawmakers arguing [[a:1154845|Tesla's largest recall yet would not fix Autopilot's failure to limit where drivers can use the feature]], and an upcoming trial over a fatal 2018 crash that may show Tesla never studied how fast drivers can retake control.
First-order effects
- Tesla's driver-alerting design stays dependent on indirect cues rather than eye tracking, leaving the company defending that choice in courtrooms and before regulators while the fatal-crash trial tests its safety claims.
- The sourcing puts internal decision-making on the record, giving plaintiffs, the NTSB, and lawmakers documentary ammunition that engineers saw the monitoring gap and leadership declined to close it.
Second-order effects
- Regulatory pressure shifts from hardware retrofits toward the operational-domain question experts keep raising — whether Tesla must restrict where Autopilot runs, not just add sensors — forcing Tesla to argue capability where critics argue scope.
- Rival autonomy programs that invested in driver monitoring gain a contrast point in the safety debate, sharpening the competitive framing around whose human-fallback assumptions hold up.
Third-order effects
- If the pattern holds, driver-assistance systems will be judged less on sensor counts than on whether makers limited deployment to conditions their monitoring could actually supervise — pushing the industry toward domain restrictions as the default safety mechanism.
- Litigation discovery is becoming the de facto audit channel for autonomous-driving design decisions, with internal engineering debates like this one shaping both liability outcomes and future regulation.
The trend: Autonomous-driving oversight is converging on the question of whether companies deploy systems where driver monitoring cannot reliably back them up, with courts and regulators filling the gap left by engineering choices made years earlier.