AI and Cities: Why and Why Now? Microsoft Chicago
A city where streetlights automatically report when they need replacing, not waiting for complaints called in from residents. A city where garbage bins and recycling are picked up when needed, not on a fixed schedule. A city where traffic of all kinds—delivery, construction, commercial, personal—is optimized in coordination with public transportation. These are all scenarios that, in one way or another, will rely on Artificial Intelligence at the city level. They are “smart cities,” to use a popular term.
People have been talking about smart cities, or at least the scenarios that they enable, for decades. Why does it seem that only now it is getting significantly more attention, more resources, and more focus? The answer lies in the technology that is needed to bring these scenarios to life has finally been realized.
The artificial intelligence that lights up these use cases can be thought of as existing in layers. Underneath Artificial Intelligence (AI), you have machine learning: algorithms that find patterns in data that would be hard (or impossible) for humans to discern. Underneath machine learning, you have data. It is at the data layer that we find the keys to the smart cities kingdom, if you will.
That data tells the stories of the city. It tells us the trends of resident experience, commerce, industry, and civic life. It tells us how key systems, from education to transit to water to energy, are working. It shines a light on where the opportunities exist for improvement. That data needs to be generated, collected, and analyzed. Two key and relatively recent advancements in technology have paved the way for the way for AI and data to help create smart cities.
The first is around data generation. We now have an enormous number of devices with cheap sensors producing massive volumes of data. Juniper research predicts that there will be over 38 billion (that’s billion with a “B”) devices connected to the internet. These devices are pumping out data at an unprecedented rate. As an example, the City of Chicago has already deployed over 100 Array of Things nodes to collect near real-time data on the city’s environment, infrastructure, and activity.
Once you have that data, you need the computing power to do something with it. In artificial intelligence, that something is “learning” with the data. This era of Internet of Things is leading to an opportunity for cities to leverage that data to create models of the systems that drive those cities. From there, they can look towards creating predictions and optimizations. That requires computing power.
Here is where the second trend comes into play: cloud computing. In the past, if you wanted to study large amounts of data, you not only had to have the storage capacity to ingest, store, and serve that data, you also had to have the computing power to analyze it. Few organizations had that kind of computing power. Fewer organizations existed to optimize the storage and computing efficiency.
That is what cloud computing is about: a computing model that can democratize massive data storage and computing power. The smallest of organizations can now access the same colossal computing assets as the largest of players. To give a sense of the scale, Microsoft spent over $7 billion in capital expenditures last year to build what is rapidly becoming the world’s leading data center infrastructure of over 100 data centers in more than 40 countries, the fundamental infrastructure of the 21st Century.
When cities bring the two together, the ability to access massive amounts of data and almost unlimited computational power, they can infuse intelligence throughout their civic systems. Cities can build models that they can use to make lives for their residents safer, healthier, more equitable, and more productive. They can use analytics to predict the impact of changes in education systems. They can use visual recognition technology to enable vehicles, bikes, and pedestrians share the public way more safely. They can use bot services to make resident interaction and engagement more natural and efficient.
In another example from Chicago, City Tech Collaborative uses the data being pumped out of city systems, combining it with data from social platforms, other sensors, as well as analytical and AI tools. This allows them to pilot solutions in the areas that every city thinks about: water, energy, transportation and congestion, and physical infrastructure. The more data and the greater the modeling capabilities, the better the insights and predictive capabilities, and the better chance they have in solving major urban issues.