Data has been the cement of smart city for the past decade. But raw data only has potential value: it requires analysis and processing. With its ever-increasing volume it became harder and harder to process. Artificial Intelligence (AI) is now a major component of any smart city program, instrumental to the optimization of any technology. AI encompasses a wide range of tools and technologies, from mere automation to understanding human language, reading images, and even adapting with deep learning technologies. All of them are already shaping what smart cities look like today and how they can evolve in the near future.
From data optimization to real-time adaptation
The race for even smarter cities means speed is more crucial than ever. With AI technology, data processing improves with greater velocity and accuracy. Data outputs and our overall ability to identify the most pressing challenges faced by any given city improved. Pattern recognition also helps predict future needs and potential human intervention related to various subjects from resilience to maintenance.
- As a consequence of global warming, floods are now more frequent in large cities such as Buenos Aires, which suffers from a particularly sensitive geographical situation combined with poor infrastructure such as pipes often obstructed with garbage. The use of IoT technology with smart sensors combined with AI technology for pattern recognition and real-time data processing allowed the city to be more reactive and avoid disasters.
- This year in DataCity Paris, the startup Fieldbox is working on a predictive maintenance device for escalators. It will be tested in Saint Lazare railway station, one of the main train hub inside Paris with over 350k people commuting or traveling through the hub every day
Complementarity or replacement?
AI-based technologies are particularly useful in the fields of energy, security and mobility.
Thanks to pattern recognition and image recognition technologies, machines can now send alerts for potential security breaches or abnormal situation, which will then be managed by security agents. Energy consumption tracking and switching from one provider to another follows the logic. But mobility is probably the most discussed field of application at the moment, as AI technology has the capacity to fully replace some human decisions:
- A simple application of AI taking over mobility rules in a city is traffic light smart control. Traffic congestion being a plague in most mega cities of the world, the ability to adjust the time a green light based on real-time traffic information could have a significant impact. This technology has already been implemented in cities like Pittsburgh where travel time was reduced by 25%.
- If traffic light turning green without a human actually controlling doesn’t seem to disturb many people, the idea of self-driving cars taking over control on the road is a different story. In 2018 63% of Americans are still afraid of riding self-driving cars. But the mindset is evolving very quickly: in 2017 78% of them were afraid of self-driving cars and in a year only of public debate, mentalities are already evolving and getting used to the idea.
AI for human-driven cities
Using AI for smart city tools and programs is not a trend, it’s a tsunami with large-scale investments from major players, the GAFA to AT&T, Oracle and a myriad of startups are ready to disrupt the field. Smart city on the whole could represent 5% of the GDP growth worldwide. But fear of machines taking over and growing concerns about data privacy — especially in Europe — should also be taken seriously.
We see this as an opportunity to also bring more transparency and dialogue between all actors, and especially engaging citizens.
One way to do it could be through education, making sure more and more people understand this technology if they have to adopt it. In July 2017, the city of Amsterdam launched a MOOC designed for marketing masses to help them understand AI better. The course is of primary use for startup founders or ecosystem members but also open to anyone interested.
Another way to use AI for better dialogue is through chatbots. Los Angeles has already implemented such technologies with Alexa Skills Kit and Chip the chatbot, training robots to answer common questions citizens may have will provide more value, and bring cities closer to their citizens.
 2015 — SNCF Gares & Connections — http://www.omnil.fr/IMG/pdf/gares_connexions_web.pdf
Originally published at datacity.numa.co on March 6, 2018.