When the topic of artificial intelligence and cars gets discussed – it’s a topic of frequent discussion online these days – it’s often in the single context of self-driving cars.
While Google is a lead in many discussions about self-driving experimentation (where a car is largely driven and controlled by a wide range of technologies), electric sports car maker Tesla gets a lot of press due to its stylish sports cars that capture imaginations with their high speeds, market-leading battery innovation, automated driving capability, and competitive prices.
Tesla’s also in the spotlight for owners of its cars getting killed in crashes where fingers point to the Autopilot technology in its cars being less than perfect for use on public roads that, combined with a driver not paying attention, is surely a recipe for lethal consequences.
Along with the ongoing discussions is lack of clarity over what people understand ‘self-driving’ to mean. In fact, there is no single definition of that phrase; rather, there is clear nuance in meaning as these definitions from the Automotive 2025: Industry without borders report (PDF) from the IBM Institute of Business Value (IBV) illustrate:
What is “self-driving”?
Automated: Driver must be present
- Partially – Driver monitors automatic functions, cannot perform non-driving tasks.
- Highly – System recognises its limitations and calls driver to take control, if needed. Driver can perform some non-driving tasks.
- Fully – System handles all situations autonomously without monitoring by driver. Driver allowed to perform non-driving tasks.
Autonomous: No driver required
- Limited – Designated areas where vehicles, infrastructure and the environment are controlled.
- Fully – Integrated with other vehicles in normal driving conditions.
Much of what we see today falls under the ‘automated’ definition. Indeed, according to the IBV report, it’s likely that we won’t see anything like full autonomous driving in the mainstream for a long time, certainly not much before 2025.
‘Automated’ is a different matter where we can expect to see considerable growth and more experimentation in the coming decade that will lead to partial mainstream use in a significant way.
Many elements are in play today, one of which – to circle back to the other conversation component that goes with ‘cars – is artificial intelligence, a notoriously tricky phrase to define with clarity that anyone can actually understand, but the definition on AlanTuring.net isn’t bad:
Artificial Intelligence (AI) is usually defined as the science of making computers do things that require intelligence when done by humans.
In the broad automotive context, AI often conjures up images of robots driving cars and lends itself very well to an emotive, scifi-ish, picture of self-driving cars. If you saw the original Total Recall movie from 1990, you’ll remember Johnny Cab (see pic at top of page).
Yet AI is present in cars today in more fundamental, practical and visible ways that are largely to do with automating many driver-related tasks – making a phone call, navigating to a destination, or avoiding a collision, for instance.
The AI ingests data from sensors and other vehicle systems and makes decisions, the outcomes of which range from asking or alerting the driver, to taking actual action to avoid or minimise danger.
This is machine learning, a subset of AI, in action.