As you’ll know by now, AI is increasingly being used to work out what jobs we get, which partners we meet and what loans we are eligible for. They are also, famously, being used to control autonomous vehicles — especially driverless cars.
One point that’s been made many times is that such vehicles will have to make moral decisions: deciding whether to kill a pensioner versus a pregnant woman, for example. The fact that most drivers have never had to make such a choice themselves is not the point. AIs with control over human lives must be encoded to make potentially life-or-death decisions, which includes choosing one undesirable option over another. However, doing this is much easier said than done, because humans find it difficult to agree what these moral values or rules should be.
Up to now, 50 sets of AI principles have been agreed internationally. However, different countries and cultures — let alone different organisations ranging from academic institutions to private companies and governments — have varying values, so no universal agreement is on the cards. For example, in the US, home of much of the AI industry, values tend to focus strongly on fairness, openness, individual rights (including privacy) and accountability. Over in China, the other major home of AI, the focus tends to be centred on group objectives and harmony. In China, the collective good is equally as important, if not more so, that individual rights. But even in China there is no universal agreement. Attempts to ‘humanise AI’ are therefore extremely difficult.
Perhaps one way forward might be to code a set of values that are better than human values, rather than being merely equal to them. Could this work? Don’t hold your breath.
Ref: Financial Times (UK) 22.7.19, ‘Formulating values for AI is hard when humans do not agree’ by J. Thornhill. www.ft.com
As Alan Tovey, industry editor of the Telegraph, has said: “If you invented the car today, it would not be allowed”. That’s probably correct. A one-ton machine capable of travelling at more than 100mph, driven by people on a public road next to other people doing the same?
Self-driving cars are one solution and the argument for them seems to be focussed upon safety. But how do you marry cars that follow rules with people that don’t? Maybe you don’t. Maybe the solution is to ban hand-steered vehicles from cities?
Roads are another problem for self-driving cars. We think of roads are being fixed, but they are not. They are changing all the time — especially with regard to potholes, debris and roadworks. And then there’s cyber. How do you ensure these cars can’t be hacked or stolen electronically?
There’s also the environmental impact. One study by the US Department of Energy found that self-driving cars could reduce fuel consumption by as much as 90 per cent — or increase it by as much as 200 per cent! One quarter of US greenhouse gas emissions come from transport, so it would be good to find out which figure is more accurate.
What is the environmental impact of replacing perfectly good cars with electric equivalents? How about the batteries? What is the environmental impact of these?
Some other unknowns include how people will use these cars, or how the value of people’s time might change as a result. Will anyone actually park a self-driving car in a city, or will the cars be instructed to drive around in circles instead? What are the implications of cars that travel much greater distances? Will self-driving cars encourage longer commutes and further urban sprawl? Where do human organs for transplant come from if people no longer die in car crashes and donate their organs?
If that hasn’t got you into a tailspin, how about some more questions? What about freedom of choice and personal autonomy? What if a car refuses to take you where you want to go, or insists it takes you via a route you don’t like?
One thing that’s certainly worth noting is that, when the car replaced the horse in the early 20th Century and solved many of the environmental issues being created by horses, people were promised a transport solution that was fast, safe and congestion free. Is the self-driving car a case of history repeating itself, especially with regard to unforeseen consequences?
One suspects that the future of the car is as much about politics and values as it is about technology.
Ref: The Economist (UK), Special report: Reinventing wheels, 3.3.18, Rand Europe, Travel in Britain in 2035 (research brief)
Remember when phones just sent text messages and made calls? Then came the apps.
Cars, which are rapidly turning into digital technology platforms, might follow a similar path. For example, cars have become Wi-Fi hotspots. Electric cars will become mobile batteries, allowing people to move energy physically from one place to another (and trade it with each other). If you are seated in a car, holding a steering wheel or looking in the rear-view mirror, the car might take your blood pressure, or monitor your eyesight or other vitals.
Best of all, perhaps, your car might dispense fresh water. A few years back there was a billboard in Lima, Peru, that condensed humid air from the coast and produced gallons of clean drinking water. A couple of engineers (Doug Martin and John Roillinger) then had the idea that a car’s air conditioning unit could do much the same thing. Given that fresh water can be rarer than cars in some remote regions this is a rather good idea, potentially.
Ref: New York Times (US) 19.10.17, ‘Cars that do much more than get you to a destination’ by B. Boudette. See also, USA Today (US) 11 January 2018, ‘Cool tech coming to a car near you’ by J. Jolly.
Forget Picassos or prime London real estate. The serious investment returns in recent years has been in classic cars, especially Ferraris.
But there are a couple of problems looming. The first is cyclical. What goes up also goes down and all it needs is another recession to start the wheels spinning in reverse. Car brands come in and out of fashion too. Not that long ago it was old Bugattis that sold for the most money, while in recent years the ten most expensive cars sold at auction have all been Ferraris. There are too many speculators in the market and arguably too many dealers and auctions too.
What about demographics and, most of all, technology? Recently, buyers have been wealthy men in their 40s, 50s, 60s and 70s reliving their bedroom wall car poster fantasies from their youth. But what happens when this market literally dies out? Do today’s young men have the same interests and passions and, if they do, might not the brands they are interested in shift gear?
More importantly, what happens to car collecting when people no longer grow up holding a driving licence or reject vehicle ownership due to environmental concerns? Most importantly of all, perhaps: What happens when driverless cars hit the roads in vast numbers?
These are all drivers of change in the years ahead.
Ref: The Deal (Aus) 8.17, ‘Gear change’ by G. Hiscock.
Along with personal jet packs, dinner in a pill and silver jumpsuits, the flying car has been a staple of science fiction visions of the future since at least the 1930s. Back in 1940, Henry Ford even said that: “Mark my word. A combination airplane and motorcar is coming. You may smile. But it is coming.”
So, where is it? The answer is: Potentially not very far away and getting closer. We’ve actually had flying cars for quite a while, largely in America. These are small planes with fold-away wings that turn into cars and vice versa. They’ve never been cheap and aren’t terribly practical either if you don’t have an airstrip in your backyard. But things are moving along with the Terrafugia Transition — recently bought by Geely, a Chinese car company — priced at £310,000. Still not cheap and still not especially practical. The 100 per cent Chinese-built Ehang 184 is a slightly different proposition. This isn’t really a car, but a large drone. So is the Kitty Hawk Flyer, a single-seat machine being developed by Larry Page, co-founder of Google.
Why is all this happening now? One reason is the emergence of new, cheap, super-lightweight materials. Another is battery technology (from other electric vehicles) plus intelligent and relatively cheap electronics for navigation and control. There’s also the speed with which drones are being developed and the public acceptance of them.
It won’t be too long before pilotless passenger drones take to the skies, especially in cities in Asia and the Middle East. Uber is one of the key companies backing such a move, which is reminiscent of the flying taxis featured in the film The Fifth Element starting Bruce Willis, as well as the ‘spinners’ in Blade Runner.
But just because something is possible doesn’t mean it will happen. Congested urban airspace is one looming problem. Another is privacy. One tech expert is quoted as saying that he’d love a flying car, but absolutely hates the idea of his neighbour having one and looking down on him.
Ref: Sunday Times magazine (UK) 28.8.18, ‘Flights of fancy’ by N. Ruffo