Can artificial intelligence out-smart cities?

Artificial intelligence (AI) is a natural fit for the smart cities movement. Just consider the billions of data points bouncing through a maze of intertwined city departments, magnified exponentially by all aspects of the Internet of Things (IoT).

It could be argued that AI will eventually be the reason cities are truly smart. However, until then, the “smartness” in many cities is more about the capturing data exhaust, rather than gaining insight from it.

The power of AI in the healthcare industry has been in its ability to aggregate billions of data points from medical cases, scholarly journals, drug trials and qualitative input from physicians and surgeons to derive  insights that humans alone could not achieve.

In this setting, success is determined by the health of the patient. In some cases, the improvement is quite dramatic; in others, it’s a matter of extending lifespan with some improvement in quality of life. Whether it’s used to reduce cost or improve treatment, AI is a patient-centric strategy.

So to what degree are cities able to get similar outcomes from AI, changing “patients” to “citizens” looking for improved quality of life in their urban environs?

One could argue that the intelligence needed to make AI work in this situation should include “citizen-generated” information, just as much as other sources. (I know many will argue that IoT feeds indirectly reflect the citizenry. But this is much like saying that my energy bills reflect how much I like cooking versus using my treadmill.)

It also begs the question of where data analytics leaves off and artificial intelligence begins. At what point does a city need to invest in cognitive computing, machine learning and AI?

One city CTO described investing in AI for basic data analytics as the equivalent of using a “chainsaw to carve your turkey.” And in tight economic times, urban taxpayers want to be sure that investments in artificial intelligence provide a tangible return.

Mayors and city councils want to be sure that AI technology can in fact outsmart the human technologists.

Most reasonably minded AI and data science professionals will tell us the biggest challenge is knowing what insights the city wants to gather in advance of the initiative. Some deployments fail because the task simply isn’t feasible, and in some cases, the investment magnifies the fact that data was totally unrelated and would likely remain so.

In other cases the resulting insight can outsmart the city. In this case, the intelligence is good but the city can’t afford to implement the recommended solution, or doesn’t have the skillsets to address the problem.

This goes back to the primary question that needs to be asked: “Will we be prepared to — and can we afford to — execute on the findings in a timely fashion?”

Nary a city official wants to explain to voters that the AI investment produced some extremely compelling findings, but unfortunately, the solutions are financially unattainable.


Will telehealth be fundamental to smart cities?

What will 2017 bring to smart cities?

How smart cities can engage citizens and bridge the Digital Divide


  1. I’ know many will argue that IoT feeds indirectly reflect the citizenry. But this is much like saying that my energy bills reflect how much I like cooking versus using my treadmill.’ Great point.

    I think this would therefore mean more involvement from citizens would help steer the change from city council to ai-assisted city management.

    Through either feedback or polling of IoT features, especially in terms of recommending what is working to optimise their day to day activities as well as queries about how long until further improvements.

    This will help AI systems better understand where problem areas are, what % of the population need the changes, and how best to deploy them. It seems to be that the more aware of differing contexts, use cases and client/citizens needs regarding service applications, the more efficient and expedient these AI systems will become.

  2. Reblogged this on Seeding concepts – diving through tech, science & society and commented:
    A short but interesting article on the challenges to be met on the way towards AI-assisting city councils in improving the management of cities

  3. I think that, in order to get recommendations from smart systems, one needs to ensure that the questions we ask or situations we analyze are sufficiently specific and atomic. The more involved your scenario, the greater the likelihood the answer or recommendations will be too broad or too expensive. Avoiding being too baroque in your requirements will likely lead to better outcomes, if only on a smaller scale.

  4. A short and delightful article. I think artificial intelligence, changing “patients” to “citizens” looking for improved quality of life in their urban environs? Good point and I would like to know Whether it’s used to reduce cost and I think in procedure to get instructions from systems, one needs to assure that the query we ask or position we figure out are adequately distinct and teeny.


  1. […] Can artificial intelligence out-smart cities? […]

  2. […] Can artificial intelligence out-smart cities? […]

  3. […] Can artificial intelligence out-smart cities? […]

  4. […] Can artificial intelligence out-smart cities? […]

Speak Your Mind


This site uses Akismet to reduce spam. Learn how your comment data is processed.