GM Product Chief: Cars Should Get Smarter, but Humanoid Robots Aren't the Answer for Building Them
Sterling Anderson, chief product officer at US automaker General Motors, explains the company's new technology roadmap built on AI, centralized computing and autonomous driving. The aim is to cut vehicle development time by more than half.
US automaker General Motors (GM) unveiled a new technology roadmap at an event in New York City last week. It centers on AI-driven vehicle development, advanced battery technology, powerful in-car computers and autonomous driving. GM's chief product officer, Sterling Anderson, discussed the plans in an interview with The Verge.
Shorter development cycles
Under the approach the company calls the "intelligence-defined vehicle," GM hopes to use artificial intelligence, centralized computing and virtualized engineering to halve its development timeline, bringing a typical four-to-five-year process down to just over two years. Anderson stresses that speed does not come at the expense of safety. In the past the company could afford only a limited number of prototypes for crash testing, which took six months; now it can run thousands of tests in a virtual environment and get feedback far sooner.
The plans come against a backdrop of strong competition from Chinese EV makers expanding rapidly around the world, as well as softer demand for electric vehicles in the US, which has forced GM to recalibrate its electrification hopes.
Autonomy and data
Anderson says safety-critical functions are processed onboard the vehicle to avoid latency and connection dropouts. Features such as the OnStar AI assistant are not safety-critical, so some of their processing happens off the vehicle.
He contrasts GM's autonomy strategy with the path taken by robotaxis: rather than limiting operation by region and speed, the system will expand road type by road type. Control will not be handed back dynamically; the system gives advance warning and, if the human does not take over, brings the car to a safe stop.
Asked about data privacy in light of earlier US Federal Trade Commission (FTC) orders over GM's sale of customer data, Anderson says end-to-end encryption means the company does not see individual data and has access only to aggregated, anonymized information.
