by GM’s New AI Assistant
General Motors is developing a new onboard artificial intelligence that, the company claims, will go even further than Google Gemini — the technology it recently integrated into its vehicles. If the mere thought of an AI chatbot in a car already makes you uneasy, get ready: GM plans to go even further.
In an interview with CNBC, GM’s director of product management for voice technology and AI/machine learning, Anna Santos, spoke about the new development:
The new native AI assistant will be “deeply integrated” and combine “conversational AI with knowledge about GM vehicles and OnStar intelligence to create capabilities that go beyond the limits of a general-purpose assistant.”
Santos added that the system will “understand the vehicle, the trip, and customer needs” better than Gemini. According to CNBC, GM is working on this technology with an unnamed large language model provider. The system may also include predictive maintenance and real-time vehicle telemetry data.
Google Gemini’s Limitations
GM’s current infotainment system runs on Google’s Android Automotive, and this year Gemini AI is being added to millions of vehicles in the 2022 model year and newer. The assistant can change climate settings, control the radio, and do everything it does on a smartphone. However, the system’s capabilities are limited: it is not a GM product and does not know the specifics of the brand’s individual models. GM’s own AI assistant can be taught that.
An AI That Really Knows Your Car
Late last year, GM’s vice president of software engineering and services, David Richardson, described the future system this way:
It will anticipate your needs, provide timely assistance, and make every trip more personalized and enjoyable.
When Can We Expect It?
Unfortunately, neither Santos nor Richardson gave any timeline for the system’s readiness. Even after development is complete, it will need to be thoroughly tested across GM’s entire lineup before being deployed.
Building its own AI makes sense for GM: the automaker has long been trying to turn cars into constantly updated digital devices. A built-in assistant with a deep understanding of a specific model could not only execute commands but also suggest actions on its own — for example, remind about maintenance, adjust a route based on driving style, or optimize EV charging.
At the same time, deploying this level of AI raises privacy concerns: collecting telemetry and analyzing driver behavior require transparent rules. Competition between Google, GM, and other automakers is also intensifying, so the race for the “smart car” is only accelerating. The success of the new system will largely depend on whether GM can ensure not only technical excellence but also user trust in data collection.
