Nevada is the first testing ground for the world’s first flying passenger drone. Chinese firm EHang will begin testing its airborne vehicle there this year. The vehicle, which can take a single person up to 100 km/h for 23 minutes, could eventually ferry people hundreds of feet aloft between one casino and another, just so long as they haven’t spent too much time at the all-you-can-eat seafood buffet.
It sounds like a utopian future, but how confident are you in autonomous vehicles?
Companies like EHang, Google, Tesla and others routinely promote the idea of driverless vehicles that navigate our cities without error, driving (or flying) better than people do. But what if the software goes wrong?
Lexus shipped a software update that effectively bricked the navigation and entertainment systems in some of its vehicles this week, causing a flood of customer complaints. Apparently one way to fix it was to disconnect and reconnect the vehicle’s battery. “Have you tried turning it off and on again?” isn’t a piece of advice you expect to have to take with a car.
Still, some companies have faith in a driverless future. U.K. insurance firm Adrian Flux has launched the country’s first ‘driverless car’ insurance policy, aimed at people with autonomous features such as self-parking in their cars. So if your vehicle does try to attack another car in a fit of artificially intelligent pique, you’ll hopefully be covered.
Google plans Skynet kill switch
Your fancy car isn’t the only thing that may require a reboot in the future. If artificial intelligence gets too uppity, we should probably have a way to turn it off. Just as well that scientists at Google and Oxford University are planning a ‘kill switch’ for AI systems that decide they don’t want to listen to us after all.
The researchers even describe a ‘big red button’ that users should be able to push to stop AI taking actions that aren’t in our best interest. They’re taking figuratively, but a real button would be awesome. Preferably with flashing lights and klaxon horns.
Hopefully, this will convince Bill Gates, Elon Musk and Stephen Hawking, who have all fretted about AI running away with itself.
Pregnancy alert! Bluetooth baby alert in effect
In South Korea, experts have created a Bluetooth-based system to alert passengers when female passengers are on the train. Women participating in the Pink Light campaign carry sensors that trigger pink lights next to priority seats when they’re nearby, meaning that people on the subway can politely give up their seats without having to actually risk talking to anyone or making eye contact. It’d probably catch on really well in Toronto and Vancouver, then.
The sensor batteries last for six months, and must be carried outside a bag for maximum signal strength. In the U.K., pregnant women wear badges on the Tube, apparently. Or — and we appreciate that this is a little avant garde — just ask if they can sit down.
Dingbat of the week
This week’s dingbat of the week award goes to Mark Zuckerberg, for having his Twitter and Pinterest accounts hacked. The Zuck’s LinkedIn password was apparently revealed as part of the massive LinkedIn breach in 2012, which were recently put up for sale. It turned out that his password was ‘dadada,’ which isn’t the strongest in the world, and was easily cracked. It sounds like he used it across various domains, which is how he got pwned.
Our question, which few people seem to be asking: how much could someone move a company’s stock price temporarily by grabbing control of the CEO’s Twitter account for a while and saying something suitably controversial?
Photo courtesty of EHang