On what should’ve been a straightforward drive in a new electric vehicle I’m testing, I found myself in an unexpected battle with its virtual assistant. While I’m keeping the car’s brand under wraps for now, the assistant’s failure to handle a simple task was hard to ignore. It was a 70-mile trip to my brother’s place, and I just needed directions. Simple, right? Well, not quite.
When I asked the assistant to guide me to ‘John Phelan’s house,’ it couldn’t do it. I tried again, but no luck. Frustrated, I let out a few choice words about its performance. To my surprise, the assistant fired back: ‘I am a virtual assistant, but your words are real. Please be respectful.’ I was taken aback. Usually, these devices just go silent, letting you think you’ve won. But this one? It had a few words of its own.
Even more puzzling was the fact that, while it couldn’t give me directions, it had no trouble calling my brother using the same contact information. This inconsistency got me thinking about the design of these AI systems. Here’s a tip for automakers: maybe skip programming assistants to lecture us on manners, especially when they’re not quite doing their job. It’s one thing to enjoy a witty exchange with a human, but getting etiquette lessons from a glitchy AI? Not so much.
This experience highlights the ongoing need to refine AI systems in vehicles. They should make driving easier, not more frustrating. Let’s hope the next generation of virtual assistants gets it right.
