Intellifusion, a Shenzhen-based AI firm that provides technology to the city's police to display the faces of jaywalkers on large LED screens at intersections, is now talking with local mobile phone carriers and social media platforms such as WeChat and Sina Weibo to develop a system where offenders will receive personal text messages as soon as they violate the rules [...] along with the fine. [...]
Facial recognition technology identifies the individual from a database and displays a photo of the jaywalking offence, the family name of the offender and part of their government identification number on large LED screens above the pavement.
In the 10 months to February this year, as many as 13,930 jaywalking offenders were recorded and displayed on the LED screen at one busy intersection in Futian district, the Shenzhen traffic police announced last month. [...]
The system will also be able to register how many times a pedestrian has violated traffic rules in the city and once this number reaches a certain level, it will affect the offender's social credit score which in turn may limit their ability to take out loans from banks, Wang said.
Previously, previously, previously, previously, previously, previously, previously, previously, previously, previously.
Is there a way to fool systems like this into thinking that I'm an automobile so that I can do whatever I want without consequences?
Whereas in the Land of the Free, this service will be outsourced to Facebook, because god hates commies.
I'm sure your China social media credit score is instantly updated with each infraction.
What are they going to do, suspend your walking license?
More like harvest your organs. The idea that China is the next big global power is a bit laughable. It's a polluted big-brother cess-pit that has run on massive amounts of dirty coal and slave labor, both of which are in increasingly short supply.
have you ever been to kentucky?
Yes, actually. It's nice but I would more seriously recommend parts of West Virginia as a vacation destination. Now is a good time of year for it.
BTW, this is all about enabling robot cars to actually work. Which is currently impossible in the USA. In "Tier 1" Chinese cities they're putting up metal barriers to make it impossible to cross the street away from cross walks.
How will robot cars answer the trolley problem, is what I want to know. Maybe one I'm in will lurch forward to stop a speeding out of control car for the greater good.
In a way I like it. Few things infuriate me more than arbitrary enforcement and the culture around that "it's ok to speed here, there are no cameras."
Wear a mask and turn the phone off. Or airplane mode, or take the battery out and bury it in layers of plastic bags and water. Then feel free to play crossy road.
But the question is, does this stuff actually work? Or has some chinese startup managed to sell their own snake oil to their tech-illiterate government, just like so many contractors in the U.S. have been doing for years?
After all, as we've seen with Cambridge Analytica, just because something doesn't work doesn't mean you can't sell it.
When asking if something "works" it's important to define goals. You're probably assuming the goal is "accurately recognize faces". That's not the case at all.
1) Someone got a fine. Maybe they say "it wasn't me". Probably lying!
2) Everyone is afraid to cross the street.
3) Performance bonuses all around!
Remember, intermittent reward and punishment leads to better compliance.
Does the CA driven manipulation really not work? You don’t need anywhere close to 100% conversion success to change the outcome of political races. In tight districts sometimes all you need to do is convince less than 1% to change their vote.
But if the change is of only 1%, how do you know whether it was thanks to Cambridge Analytica? Maybe it was the weather, or maybe it was because of that one TV ad, or maybe it's simply that Hillary didn't campaign enough in Wisconsin.
If there was a 1% swing in the results and Cambridge Analytica took credit for it, all that proves is that they are like pretty much every PR agency and marketer in the history of the discipline.
I’m not a statistician and can’t give you a confident answer though am guessing that there’s some rigorous analysis that can quantify a probability of conversion “good enough” to be worth the effort to try.
BTW what Cambridge Analytica does isn’t really new. They just lowered the cost per influence contact by using the internet and automation. When I volunteered to canvas for political campaigns in college we did essentially the same thing except with pencil and paper and knocking on doors. Analysis from polls identified the tight races and which districts had the most voters on the fence. We even had data on which addresses were likely already decided either positive or negative so we knew not to waste time there. It wasn’t an exact science back then either, just a way to focus efforts as efficiently as possible.
Any time you find yourself saying, "it isn't really new, it just" you should probably stop there.
The Internet isn't really new, it's just a telegraph, but faster and cheaper.
The traditional meat driven political machine scaled only as big as you could find volunteers sympathetic to your cause. This new method requires no sympathy from the computers canvassing screens and scales as big as money behind the cause. A few billionaires sharing a common philosophy could steer the course of a society. Hope their intentions are good and sincere.
This is messed up :P