For the past 12 months, Chinese e-commerce giant Alibaba has been slurping up video feeds, social media data, traffic information and other data from Hangzhou city for its City Brain project. The stated goal was to improve life in Hangzhou by letting artificial intelligence process this data and use it to control aspects of urban life. It seems to have worked. The trial has been so successful that the company is now packaging the system for export to other places in China 鈥� and eventually the rest of the world.
Using AI to optimise Hangzhou has had many positive effects. Traffic congestion is down, road accidents are automatically detected and responded to faster, and illegal parking is tracked in real time. If someone breaks the law, they too can be tracked throughout the city before being picked up by the police. 鈥淚n China, people have less concern with privacy, which allows us to move faster,鈥� said Xian-Sheng Hua, who manages AI at Alibaba, speaking at .
Consider the example of traffic flow. Using hundreds of thousands of cameras dotted throughout the city, Alibaba can track almost every car on every road. It can instantly detect crashes, blockages or parking violations, and automatically notify the police to deal with them. The system can predict the traffic flow 10 minutes ahead of time with 90 per cent accuracy, and responds by changing traffic light patterns to even out congestion. It can even send text messages to people to help them plan different routes. 鈥淐ity Brain is about comprehensive cognition,鈥� said Hua.
Advertisement
But some see clear downsides. 鈥淎 fully 鈥榮mart鈥� city means that pretty much every aspect of your life is tracked,鈥� says Paul Bernal at the University of East Anglia, UK.
鈥淣ormal鈥� behaviour
It鈥檚 a conundrum. It鈥檚 easy to see the attraction of a super-smart city: in Hangzhou, the average journey time has improved by 10 per cent and better traffic flow means less fuel usage, which is better for air quality and better for the planet. Imagine doing the same for the electricity grid or the water supply and the benefits are multiplied.
But Bernal cautions that this has to be weighed against the potential cost. 鈥淭his allows a great deal of efficiency, but the privacy issues are huge,鈥� he says. His point is more apparent in the city鈥檚 other optimisations. Alibaba鈥檚 algorithms index all video footage so that it can be searched quickly later. If police give the City Brain a picture of, say, a motorcyclist, the AI will return exact matches detailing where and when that motorcyclist has been in the city since the project began.
鈥淚t鈥檚 easy to identify when people are not following the 鈥榥ormal鈥� behaviour patterns. Having identified people who are not 鈥榥ormal鈥�, they can of course then be tracked聽鈥� and who they meet with, where they go and so forth can also be quickly identified,鈥� says Bernal. 鈥淎s a way to control dissident movements or anything the authorities don鈥檛 like, it鈥檚 perfect.鈥�
There have been reports worldwide of police misusing technology for everything from looking up people they are interested in, to tracking their exes. Only a few months ago, a policeman in the UK admitted to using a police helicopter鈥檚 camera to film his 鈥渟winging friends鈥� having sex in their garden. People with power often abuse it. 鈥淭his kind of system would allow this to happen to a much more extreme degree,鈥� says Bernal.
Smart city spread
Abuse or not, becoming overly reliant on the City Brain might introduce mistakes. There are imperfections in the City Brain algorithms. Difficulties come from the huge variety in cameras: some produce footage that is shaky or of inferior quality, and others are better but further away from the search object. Stitching together reliable narratives from such a wide variation in image quality is tricky. Still, Hua argues that it can provide a starting point for further investigation. 鈥淲e can also do people search, but it鈥檚 very difficult. Currently we make many more mistakes here,鈥� says Hua.
It will be up to other local authorities to weigh up the pros and cons; Alibaba is turning the City Brain into a 鈥減op-up鈥� product other cities can adopt. It is already in talks with several in China, including Shanghai.
And similar projects are starting up elsewhere. The Canadian city of Toronto recently announced plans to turn one region into an interconnected smart hub with the help of Sidewalk Labs, a subsidiary of Google鈥檚 parent company Alphabet. It鈥檚 just one of many smart city projects under way across the world. However, the Hangzhou City Brain project is in another league. The sheer amount of data it uses pushes the boundaries of what it means to call something a smart city.
