Artificial intelligence has long been a buzz-word in industries as varied as banking to gaming, but the popular conception of a machine you can talk to has obscured the fact that many technologies already use some version of artificial intelligence in their products. Take the automotive industry, for example: while the obvious and likely near iteration of AI might be seen in the implementation of self-driving cars, it has already found tried and tested uses in many other aspects of the sector. The elements that might make a self-driving car are already in place in many modern vehicles, and most of them rely on some aspect of AI to function reliably and safely.
One of the closest technologies to a self-driving vehicle is the ability to enable assisted driving. In terms of usage, in many ways it’s not that different from technologies like cruise control and anti-lock brakes – it’s a technology designed to make a driver’s workload easier, rather than remove it completely. Behind the scenes, things get a lot more complicated. The ability to recognise and react to other road users, weather and terrain conditions, and of course unexpected events, is a decision-making process orders of magnitude beyond what could have been attempted twenty years ago. It requires powerful AI with the ability to respond to situations faster, and make decisions better, than a human probably would in the same situation.
The fundamental difference between this sort of technology and more simple software solutions lies in its ability to measure and understand situations that is has no previous direct experience with. Because the number of variables in most traffic situations are so large, it would be almost impossible to produce a program that had predetermined outcomes for every possible situation. AI solves this problem by providing a set of rules based around situational awareness instead – a faster, more efficient way to control complex systems.
That situational awareness is provided in part by another use of AI – image recognition. To human perception, the relationship between what we see and knowing what that thing is, is inherent to the way our minds work. Not so for computers – at least before AI with learning algorithms became widespread. With current machine learning technology, computers are becoming far more able to do something we take for granted – know that if it looks like a car, and it acts like a car, then it probably is a car. Without that simple-seaming leap, everything from assisted driving to fully autonomous vehicles become a far more difficult problem to solve.
This technology is becoming widespread in the automotive industry, and that trend can only increase. With the continued efficient use of AI, cars will become safer, more comfortable, and cheaper to run, with fuel costs and insurance costs dropping. It’s also a technology that will improve rapidly, with software updates becoming a routine part of servicing. When you think about how rapidly other consumer technologies have developed in recent year, it’s an exciting time to be driving, and looks like staying that way for the foreseeable future.