Photo: HP Mega Trends
In the past decade or so, we’ve seen the rate of technology development gaining an overwhelming speed, with important new discoveries that will have enormous impact on the world. Like artificial intelligence.
But that wasn’t always the case. For example in the 1970s, artificial intelligence existed only in science fiction novels, and some movies here and there. Travel to about 30 years ago and the concept of AI was already taking shape, but the success rate was not that impressive.
Around the mid 2000’s there was a substantial breakthrough in machine learning, which made the concept of Deep Learning become feasible, which in turn became the Big Bang for artificial intelligence, aka AI. Fast forward to today, and we have plenty of AI neuronal networks that help us process digital information such as browsing, text translation or even the photos we take with our smartphones.
There’s one area though where AI promises to open up exciting new opportunities, and that is the automotive industry. But how will that actually change the future of our cars?
The first example that comes to mind is autonomous driving, and while cars that attempt to offer such privileges have been actively developed over the past decade, and with substantial progress, the end result is far from perfect. So for us petrol-heads, it’s a bit of a relief knowing that in the next five to ten years, artificial intelligence will hardly be able to replace, or at least seriously threaten, the necessity of an experienced driver.
However, there’s a popular belief that trying to bring artificial intelligence at a human level is a futile, as there would be other applications and technologies that should be developed instead of trying to recreate human intelligence.
To quote Peter Norvig, the author of Artificial Intelligence: A Modern Approach: “We know how to build real intelligence - my wife and I did it twice, (...). We don’t need to duplicate humans. That’s why I focus on having tools to help us, rather than duplicate what we already know how to do. We want humans and machines to partner and do something that they cannot do on their own.”
This process would create a new breed, call it an “artificial driver”, where the AI in the car will not only observe when you exceed speed limits, park the car or stop you at intersections, but also be able to resolve ethical issues at a higher speed than we can. For example, and an unfortunate one for that matter, when an accident is inevitable and the AI has to choose between two bad options.
For the sake of this exercise, imagine you are driving a car and all of a sudden the brakes fail, and on the road ahead there are 2 people crossing the street. You could steer in a different direction, but that way has a person crossing the street as well. It’s a situation where you could unintentionally run over 2 people, versus intentionally running over just one. Not ideal, I know.
This moral dilemma is known as the Trolley Problem, and it in this particular scenario, how a vehicle with artificial intelligence could make a choice in a situation where humans are overwhelmed. There are no definite answers to such questions though, and we’ll probably have to wait some more for a system that will have the “right answers”.
Nevertheless, MIT has taken the initiative and created a website that offers everyone the opportunity to solve such moral dilemmas for the sake of automotive AI development. The end result would be that the data gathered from people would help developers “teach” artificial intelligence to make ethical decisions, guided by human indicators, at a much higher speed.
If we consider the existing innovative AI technologies that we have today, we can see that they are closely linked with the goal of reproducing the human mind, in a digital form. Neural networks are being named and developed following the structure of the human brain, and machine learning is based on experience and repetition, mainly like the ways of teaching people.
To paint a broader picture, the German scientist who invented the electroencephalogram (EEG) actually wanted to develop a telepathy machine, instead he brought us the most important tool for studying the human brain.
That being said, automotive AI should not focus solely on autonomous driving, and if it does, maybe we’ll end up with a bridge for creating a synergy between man and machine, which will better solve complex scenarios that could occur on the road.
Around the mid 2000’s there was a substantial breakthrough in machine learning, which made the concept of Deep Learning become feasible, which in turn became the Big Bang for artificial intelligence, aka AI. Fast forward to today, and we have plenty of AI neuronal networks that help us process digital information such as browsing, text translation or even the photos we take with our smartphones.
There’s one area though where AI promises to open up exciting new opportunities, and that is the automotive industry. But how will that actually change the future of our cars?
The first example that comes to mind is autonomous driving, and while cars that attempt to offer such privileges have been actively developed over the past decade, and with substantial progress, the end result is far from perfect. So for us petrol-heads, it’s a bit of a relief knowing that in the next five to ten years, artificial intelligence will hardly be able to replace, or at least seriously threaten, the necessity of an experienced driver.
However, there’s a popular belief that trying to bring artificial intelligence at a human level is a futile, as there would be other applications and technologies that should be developed instead of trying to recreate human intelligence.
To quote Peter Norvig, the author of Artificial Intelligence: A Modern Approach: “We know how to build real intelligence - my wife and I did it twice, (...). We don’t need to duplicate humans. That’s why I focus on having tools to help us, rather than duplicate what we already know how to do. We want humans and machines to partner and do something that they cannot do on their own.”
This process would create a new breed, call it an “artificial driver”, where the AI in the car will not only observe when you exceed speed limits, park the car or stop you at intersections, but also be able to resolve ethical issues at a higher speed than we can. For example, and an unfortunate one for that matter, when an accident is inevitable and the AI has to choose between two bad options.
For the sake of this exercise, imagine you are driving a car and all of a sudden the brakes fail, and on the road ahead there are 2 people crossing the street. You could steer in a different direction, but that way has a person crossing the street as well. It’s a situation where you could unintentionally run over 2 people, versus intentionally running over just one. Not ideal, I know.
This moral dilemma is known as the Trolley Problem, and it in this particular scenario, how a vehicle with artificial intelligence could make a choice in a situation where humans are overwhelmed. There are no definite answers to such questions though, and we’ll probably have to wait some more for a system that will have the “right answers”.
Nevertheless, MIT has taken the initiative and created a website that offers everyone the opportunity to solve such moral dilemmas for the sake of automotive AI development. The end result would be that the data gathered from people would help developers “teach” artificial intelligence to make ethical decisions, guided by human indicators, at a much higher speed.
If we consider the existing innovative AI technologies that we have today, we can see that they are closely linked with the goal of reproducing the human mind, in a digital form. Neural networks are being named and developed following the structure of the human brain, and machine learning is based on experience and repetition, mainly like the ways of teaching people.
To paint a broader picture, the German scientist who invented the electroencephalogram (EEG) actually wanted to develop a telepathy machine, instead he brought us the most important tool for studying the human brain.
That being said, automotive AI should not focus solely on autonomous driving, and if it does, maybe we’ll end up with a bridge for creating a synergy between man and machine, which will better solve complex scenarios that could occur on the road.
Google News