Targeting AI: Tesla undervalues self-driving safety tech
The top electric car maker comes under heavy criticism for not investing in multiple sensors for its full self-driving system to improve driver and pedestrian safety.
Sam Abuelsamid thinks Tesla's self-driving technology is unsafe.
The mobility ecosystem analyst at Guidehouse Insights is a vocal critic of the electric vehicle giant's AI-powered "Autopilot" driver assist and full-self driving technology.
A former mechanical engineer, automotive journalist, and Ford and General Motors employee, Abuelsamid also charges that the National Highway Traffic Safety Administration (NHTSA) has grossly undervalued safety considerations for self-driving and partially self-driving vehicles.
While Abuelsamid acknowledges that Tesla has advanced driving technology by appealing to consumers and popularizing electric vehicles, he also refuses to call such vehicles "autonomous." Instead, he refers to them as "automated" because, as he points out, few fully driverless vehicles are on the road.
In addition, Abuelsamid contends that Tesla has tried to do safety "on the cheap" by relying on cameras only to power Autopilot features and not using considerably more expensive sensor arrays.
"I think they've been utterly reckless and irresponsible in their approach to automated driving by putting experimental software in the hands of average consumers who are not trained in how to properly test and evaluate this kind of safety critical software," Abuelsamid said on the Targeting AI podcast.
Sam AbuelsamidAnalyst, Guidehouse Insights
Meanwhile, autonomous vehicle technology vendors, including Cruise, Waymo, Zoox and Motional, are using multiple types of sensors, he says.
One prominent Tesla fan and investor, Ross Gerber, CEO of Gerber Kawasaki Wealth and Investment Management, has disputed Tesla safety critics. He argues that autonomously driven Teslas will get increasingly safer with hundreds of thousands of consumers driving and testing out the beta version of the popular carmaker's full self-driving capability.
But Abuelsamid also faults NHTSA for failing to effectively oversee safety aspects of autonomous vehicle technology vendors.
"I think the National Highway Traffic Safety Administration has been negligent in not doing more to require sharing of data from these test vehicles to build an understanding of how these things function," he says. "At a minimum what we need is the electronic equivalent of what we have to do as humans to get a driver's license."
Go to TechTarget News for reports on autonomous vehicle technology and other AI developments.
Shaun Sutner is senior news director for TechTarget Editorial's enterprise AI, business analytics, data management, customer experience and unified communications coverage areas.
Esther Ajao is a TechTarget news writer covering artificial intelligence software and systems.
Together, they host the "Targeting AI" podcast series.