City procurement practices could hamper AI deployment

The market for artificial intelligence (AI) solutions in cities is expected to see “substantial” growth in the coming decade – particularly for use cases related to sustainability, according to Guidehouse Insights.

A new report looks at the market for AI and predictive analytics platforms for transportation, energy management and environmental quality in cities.

The consulting firm expects annual sales revenue for AI focused on city sustainability to grow from US$693.3 million in 2023 to US$6.5 billion by 2032, but the report also highlights market barriers including municipal procurement practices, and privacy and cybersecurity concerns.

“In the coming decade, as climate challenges increase, deployment of data-collecting IoT devices expands, and cities become familiar with the outcomes that can be achieved through AI systems, the market is expected to grow substantially,” said Grant Samms, research analyst with Guidehouse Insights. “AI’s ability to process large amounts of data from diverse sources makes it a useful tool in addressing difficult, sticky problems, which have many interacting factors that may be difficult to detect manually.”

Recent infrastructure and climate investment policies in the US and Europe are expected to further boost the AI market. These include the Infrastructure Investment and Jobs Act and the Inflation Reduction Act in the US. Through a European Commission programme, over 100 cities have pledged to become climate-neutral by 2030.

Growth areas

In the three major sustainability AI segments – transportation, energy management and environmental quality – the report says that transportation and energy management have found early success thanks to “clearer paths for return on investment”, large existing data sets, and widespread deployment of smart devices such as intersection cameras and advanced metering infrastructure (AMI) energy meters.

Samms told Cities Today: “While AI is an incredibly adaptable tool and its possible applications many, the one that stands out to me currently is the use of visual AI for transportation applications. Algorithms can be trained to take live video and translate it into all manner of data that is more voluminous, more accurate and less expensive than traditional data collection methods.

“With this data in hand, cities can make much more nuanced decisions about their operations and address problems that are difficult to solve.

“For example, the city of Peterborough, Ontario used visual AI at intersections to study traffic patterns, alter traffic signal timing, and ultimately reduce travel time on a major roadway by upwards of 30 percent. Similar AI can be used to help plan micromobility projects, improve pedestrian safety, and quickly map road surface conditions.”

Challenges

According to Guidehouse, cost is not typically the largest obstacle to AI deployment. Rather, the report finds that one of the largest barriers is municipal procurement practices, which are not designed to accommodate many AI companies’ pricing models.

Samms explained: “Many cities are used to budgeting projects as discrete expenditures. For instance, spending a pot of money today for a product that will last 10 years before needing to be replaced.”

AI applications are often priced in different ways, such as by volume of data processed or ‘as-a-service’, with the price fluctuating as the features a customer needs change.

“This approach grants more flexibility but blends OpEx and CapEx in ways that cities aren’t traditionally used to considering,” said Samms. “Additionally, many of these technologies are coming to market from younger start-up companies that some cities are nervous to operate with out of fear that the company may not be around for the full contract term. They may impose requirements on companies that smaller start-ups simply cannot meet.”

Cities must also address concerns about potential bias in AI systems and how people’s privacy will be protected, as well as managing the IT and cybersecurity demands of new AI technologies.

There is a growing movement in some cities to demonstrate transparency in how data and algorithms are used – including through sensor signage and public algorithm registers.

Frontrunners have highlighted challenges, though, such as the staff time commitment required, defining what counts as an algorithm, rooting out all the applications already being used, and restrictive vendor confidentiality clauses.

The Atlas for Urban AI, launched last year by the Cities Coalition for Digital Rights and others, details AI projects across the globe, with a focus on alignment with ethical principles.