Inside the rise of software-defined vehicles

FROM THINGS LIKE COCA-COLA’S ‘AI-created flavour’ to the Swan AI-powered kettle, the modern-day recipe for product innovation seems to increasingly be: take existing product, add artificial intelligence. Bon appétit! While we may find ourselves questioning the real-life use case of many of these products, improving the conversational AI found in our personal vehicles is becoming a necessity. As anyone who has attempted to use in-car conversational AI will know, compared to some of our home products and personal devices, there’s still a clunky gap to fill. As we move towards the future of the software-defined vehicle (SDV), where software will control everything in the car and connect it to the internet of things, the personal car will increasingly become an extension of our home and work lives, rather than just a vehicle to get us places.

Granted, you might be reading this from your inner-city dwelling, thinking, ‘But I don’t commute in my car that much’. Yet, recent statistics show more people are leaving large cities for regional areas; therefore, futurists are starting to predict a new cohort of future car buyers, cementing the vehicle as a third space.

Much of the clunkiness within existing in-car AI systems has to do with the lag between automotive product delivery and tech industry developments—there’s a big gap in time between developing, delivering and updating a car, versus an update that can be plugged into a website. This is one of the reasons over-the-air-updates have been a game changer for the automotive industry—as new software is rolled out, some cars with the ability to update wirelessly will gain the newest operating systems and patches, meaning you don’t need the hottest, newest, latest model to benefit (yes, just like your phone). Earlier this year, a handful of original equipment manufacturers (OEMs) announced fresh partnerships with generative AI and Large Language Models (LLMs–AI algorithms trained on huge data sets), giving us a taste of what’s to come. BMW is building a new and more human-like conversational assistant, thanks to developments with Amazon’s Alexa, which has been available in its cars since 2018. Volkswagen and Sköda, meanwhile, have plans to introduce ChatGPT into their cars and will use the software to offer more intuitive and ‘conversational functions’, which could include navigation, climate, vehicle-specific information and general knowledge questions.

Mercedes-Benz, which already brags one of the most intuitive systems, has been testing various LLM software for its next-gen MBUX visual assistant, and claims it will be able to stand up next to leading generative AI models, will be more emotional, human-like and will have the ability to address complex, realistic questions and prompts. As a sign of the times, Mercedes-Benz is also adding more in-car videoconferencing and streaming software, including Microsoft Teams and Sony Pictures Entertainment’s RIDEVU, while others are looking at how to incorporate more gaming platforms.

Unlike Coke’s AI-created soft drinks, the future of generative AI within an automotive context has plenty of use cases beyond the obvious. Imagine being able to just ask your car what that annoying beep is, when it thinks you’ll need certain components serviced, or how many km you’ll need of range/fuel based on your schedule that day. Maybe around dinner time, it might remind you that you’re close to a highly rated restaurant of your favourite cuisine. Perhaps it’ll help you write emails, memos and messages to your coworkers, based on specific prompts. Or even, looking down the line towards a future of connected cars, it might even be able to let you know when a friend is nearby. Seeing as though safe, fully autonomous vehicles are proving to be more of a pipe dream these days, importantly, you’ll be able to do all of this while keeping your hands on the wheel and eyes on the road.