Despite some remarkable advances in technology over recent years with self-driving cars, you’ll have to keep your eyes on the road, your hands upon the wheel, and your foot near the brake for the foreseeable future.
A story in The Daily Beast by Ogi Ogas and Sai Gaddam points out that while the “artificial minds” of autonomous cars are making great leaps forward in intelligence, they lack one crucial factor: consciousness.
The authors, who are researchers from Boston University, point out that even the most intelligent cars are stuck at the intellectual level of an ant, and lacking consciousness.
Consciousness is specific to vertebrates, animals with a backbone, and allows them to make autonomous decisions in real time, in a complex and unpredictable environment that’s constantly changing.
You can see it in any vertebrate; and consciousness reaches its highest form in human beings.
The authors use the example of Indian roads, which are a chaos of vehicles, people working on the side of the road, cars honking constantly, animals darting everywhere, and motorbikes swerving in and out.
To break free of human minders and a reliance on stable, familiar environments, a self-driving car must be able to evaluate and respond to multiple simultaneous context-dependent inputs (including novel and unique inputs) from different modalities (audio, visual, lidar, navigational) arising from every direction (front, back, left, right, above, below) while pursuing an objective (drive to Bazar Road for a delivery).
On Indian roads it's a task that’s well beyond the ability of today’s self-driving vehicles.
The development of self-driving vehicles requires Deep Learning algorithms based on events that are statistically likely to happen on the roads to build a model of a world that’s likely to happen in the future.
A lot of the time that might work quite well, but the problem comes with unexpected events, and that’s where consciousness becomes irreplaceable.
Consciousness allows you to make sense of the information streaming into your brain from all of your various senses; and to prioritise them for the right response.
It’s what lets you react immediately to unique and novel situations.
To take an extreme example, think of a light aircraft with engine trouble flying overhead while you’re driving down the highway.
The pilot needs to land as fast as possible, and the road is the least worst option for them.
So, there you are, driving down the road, when suddenly there’s a light aircraft touching down in front of you.
You’ve certainly never seen that before, but nevertheless you’ve processed the information, hit the brakes, and taken evasive action.
But it’s a fair bet that situation has never been programmed into an autonomous vehicle; because it’s so unlikely to happen, and how it reacts is anyone’s guess. Perhaps it interprets it as a big bird that’s likely to swoop away. Who knows? But crashing into that plane is pretty likely while it’s making up its “mind.”
This is an extreme example of course, but novel situations occur on our roads every day, like ambiguous lane markings during road works, overturned vehicles, and spilt loads, all situations requiring a conscious response that could confuse an autonomous vehicle but which we could handle easily.
So, sit up straight and pay attention on the roads, because it looks like you’ll still be driving for quite a while yet.