Robots on the Road: Should Riders Trust AI Behind the Wheel?
As Bosch and Volkswagen push AI-driven cars closer, bikers are left wondering if robots can really be trusted to spot, predict, and protect us.
Autonomous or self-driving cars. We’ve been promised them for years now, usually accompanied by slick press shots of a happy couple watching Netflix while their car silently drives them to dinner.
In reality? They’re still a long way off. Cases of phantom braking, unpredictable interventions, and the occasional confused machine slamming on the anchors in the middle of a motorway show that the dream is, for now, still exactly that: a dream.
But that hasn’t stopped manufacturers (and governments) from ploughing billions into the technology. Bosch and CARIAD (Volkswagen’s software arm) are the latest to double down, working together under the Automated Driving Alliance banner. Their aim? A shiny new AI-based software stack that promises Level 2 and Level 3 driving assistance by 2026 - whatever that means.
On paper, it sounds impressive. Using generative AI techniques, the system can analyse complex urban traffic, predict what other road users might do, and even combine language and vision cues in a way that mimics human reasoning. It’s being tested across Europe, Japan, and the United States, and the hope that it can cope with local quirks. The grand plan is to roll it into Volkswagen’s next vehicle platform, with Bosch offering the tech to pretty much anyone who’ll buy it.
Sounds clever. But if you’re on two wheels, like me, it’s hard not to feel a twinge of unease.
When AI meets motorcycles
The official line is that all of this is about safety. Bosch, now positioning itself as much a software company as a hardware giant, is also working on “vehicle motion management” - essentially a clever brain that ties together steering, braking, powertrain and chassis control. It is supposed to mean that the car can make smoother, safer interventions when something unexpected happens. Hazard warnings, road friction estimation, and even remote health monitoring of vehicle components are all part of the package.
From a driver’s perspective, brilliant. From a biker’s? Well, it raises more questions than answers.
Motorcyclists are the ultimate edge case in traffic. We’re small, quick, and capable of manoeuvres that baffle the logic of even seasoned drivers. Filtering through traffic, accelerating into a safe gap, or simply leaning into a corner are all things that don’t fit neatly into the car driving rulebook. Can a robot car really predict that a GSX-R is about to overtake three cars in one go, or that a delivery rider will squeeze up the inside of a bus?
Phantom braking in a car is an inconvenience. Phantom braking when you’ve got a biker tucked in behind you? That’s potentially lethal. And we aren’t sensationalising here. This kind of thing has happened. There was a very similar scenario in America, in that a self-driving car ‘saw’ a motorcycle further up the road, but because it didn’t recognise the bike as a bike, it ploughed into the back of it killing the rider.
The makers of that system quickly did a software update and proclaimed it was fixed. Sadly for the rider, that’s too late.
The 2026 deadline
Volkswagen and Bosch say the production-ready system will land in mid-2026. By then, they reckon, AI will be sophisticated enough to not just see a motorbike, but to understand what the rider might do next. Big claims, especially when you consider that even the sharpest human drivers often struggle to anticipate us.
The companies talk about “mimicking human reasoning” using Vision-Language-Action models. Which is all well and good, but riders don’t always act rationally. We brake late, we accelerate hard, and we sometimes take lines through traffic that a computer simply can’t fathom. Can an algorithm, no matter how clever, ever truly understand the motorcycling mindset?
What it means for us
For bikers, the worry isn’t just whether the tech works in lab conditions, it’s how it behaves in the messy, unpredictable world of real traffic. Will a robot car nudge into a junction because it’s “predicted” we’ll stop, when in reality we’ve already committed to going? Will it hesitate when we filter, blocking the gap? Or worse, will it slam the brakes on because it “thought” it saw something? You’d hope that, as Bosch also creates advanced driver assistance systems (ADAS) for bikes, it would take this into account.
And I'm emphasising the ‘hope’, there.
Bosch and Volkswagen promise that AI will make the roads safer for everyone. And maybe one day it will. But until these systems prove they can cope with the split-second unpredictability of two wheels, I’ll be giving any robot-driven car a wide berth. Because while the future might belong to AI, the present still belongs to us, and I’d quite like to survive long enough to enjoy it.
Find the latest motorcycle news on Visordown.com
Simon (AKA Toad) has been working within the motorcycle industry for more than twelve years, and a member of the motorcyclist family since birth. Growing up around British-built single and twins, it was inevitable that a life on two wheels was on the cards, from a very early age!
