General Motors has introduced a new technology roadmap focused on AI-driven vehicle development, advanced battery technology, and autonomous driving systems. The company aims to utilize centralized computing and virtualized engineering to reduce vehicle development timelines from four to five years down to just over two years.
The automaker is integrating transformer-based foundation models into its design processes, leveraging internal data from decades of engineering files. For safety-critical functions, GM mandates that AI inference occurs onboard the vehicle to avoid latency issues, while non-critical features like OnStar AI interactions may utilize cloud processing.
GM is implementing a common electrical system and centralized compute architecture across its entire vehicle portfolio to eliminate software complexity. This strategy allows software engineers to develop for a unified platform rather than supporting multiple disparate vehicle configurations.
Regarding autonomous driving, GM plans to deploy systems based on road-type specific settings rather than the regionally bounded approach used by robotaxi operators. The company intends to manage human-in-the-loop handoffs by providing advance notice to drivers, ensuring the vehicle reaches a safe state if the driver does not take control.
To address data privacy concerns, GM states it is implementing end-to-end encryption for user data. The company claims that individual data remains under user control, while information accessible to GM is aggregated and anonymized. Users can revoke permissions or delete their data at any time.
