AI-enabled Smart Cities: What to Get Right

In October 2017, Alphabet’s subsidiary Sidewalk Labs and the Canadian government teamed up to create a new age smart city. Google’s sister company was tasked to design a modern, AI-enabled city on Toronto’s former industrial waterfront.

Named Quayside, the place was slated to have an underground garbage disposal, heated sidewalks, and skyscrapers made from timber. Sensors in the city would monitor the area and provide data to AI-powered devices, which would direct traffic and manage stormwater.

It was hyped up to be the city of the future but eventually died quietly in the May of last year.

Sidewalk Labs Chief Executive Officer Dan Doctoroff attributed Quayside’s cancellation to the disruption in real estate caused by COVID-19. However, the project was facing issues long before the outbreak.

Quayside failed to build public trust. The terms of the agreement between the stakeholders were kept secret, inviting more skepticism from activists and community groups. Privacy concerns were also raised as the lack of transparency around the city’s data handling from residents and visitors was criticized.

Although Alphabet’s ambitious project will never see the light of day —its failure has left plenty of lessons for those who wish to design smart cities.

The prospect of smart cities is exciting. Emerging technologies are reducing the cost of doing business, disrupting healthcare, and elevating every facet of human life. Cities administered through algorithms are a part of the natural progression. But for smart cities to fare better than Quayside, their designers need to understand that cities have different layers.

Data is the foundation of smart cities - it provides insights that help in creating a smooth lifestyle. However, to deliver the right solutions, planners must establish sustainable data and technology policies.

Here are few things to consider while planning a smart city:

1. Build Trust from the Get-Go

Smart cities cannot function if all stakeholders involved don’t have a certain trust in them. Residents should be assured that their data is secured and is only being used in accordance with their wishes. Since data is the lifeline of the city, policies should be designed to balance utility and privacy.

Planners should not expect citizens to completely give up their private information in return for superior services. Transparency should be at the center of policies, legislation, and technologies that govern the city.

Planners need to ensure that smart cities don’t turn into mass surveillance cities. One way to do this is by differentiating de-identifiable data from personally identified data. Personally identifiable data collected from smart city sensors shouldn’t be shared freely. Both public and private stakeholders should enforce mechanisms to protect against the abuse of the data collected.

Unfortunately, even Artificial Intelligence can be biased based on race, ethnicity, and gender. This happens when the dataset used to train the algorithms is non-diverse and over-represents a certain community.

When these algorithms are used for crime prevention, it creates a dangerous situation where certain races and genders face discrimination. Different studies have shown that predictive policing tools perpetuate systemic racism.

For AI to be unbiased, the data needs to be impartial, and the team of developers should be from diverse backgrounds. The stakeholders should also bring a level of transparency and allow third parties to audit the dataset.

However, even this is not enough. The role of AI in predictive policing and law enforcement should be limited, as there can always be human errors while designing such algorithms.

3. Focus on the Solutions, Not the Tech Itself

While AI, IoT, and big data can enable smart cities to provide solutions for everything from traffic management to security, the focus should remain on the outcomes and not technology.

At the end of the day, these cities are about creating better results for businesses, administrators, and residents alike. Solutions should be created based on the city’s needs, be it the health and safety of the public or the economic opportunities available to them.

4. Make the City Inclusive by Being Proactive

By nature, smart cities prioritize tech-savvy individuals over those who are not well-acquainted with advanced technologies. Since the latter group is not benefiting from the city's innovative service, it’s highly unlikely they would be involved in the decision-making process. This creates a dangerous divide between different segments of society.

IBM’s Everyday Ethics for Artificial Intelligence guide singles out explainability as one of the important aspects of ethical AI.

Smart systems governing the city should be an important part of the academic course. Educational material in the form of podcasts, videos, and fliers should be easily available. Citizen’s participation should be simplified feedback from citizens with diverse backgrounds should be gathered and processed to make necessary adjustments.

5. Leverage Public-Private Partnership

Whether they are being integrated or built from scratch, smart-city technologies assist administrators in getting more out of the city’s assets. These technologies can introduce new capabilities to optimize maintenance and utilize the city’s physical assets to their fullest potential.

Conventionally, it takes extremely long and capital-intensive infrastructure investment to upgrade the cities. But smart solutions combined with traditional construction can streamline the process and allow cities to better respond to growing demands.

For instance, Copenhagen’s LOOP City will have autonomous minibusses running alongside the planned light rail. These buses will be managed by private investors and help solve traffic congestion.

But private participation should be limited to businesses. Advocacy groups representing disadvantaged communities should be involved every step of the way.

Smart cities function on real-time data. They collect information as residents move around the city, using tools such as facial recognition software, traffic cameras, and license plate readers, among others.

Data collected can create solutions for mobility, public safety and also reduce the overall administration costs. But smart cities need to address concerns whether they’re related to mass surveillance, racial bias, or inclusivity. Otherwise, the cities of the future might never materialize.