Skip to main content

Three new ways AI is helping cities to become smarter

Submitted by kevin ebi on August 16, 2018

Scientists have been talking about artificial intelligence (AI) for decades, but in some cities, the future is already here. Some smart systems are already optimizing themselves and the benefits are real: reduced risk in public infrastructure projects, streamlined property appraisals, faster traffic.

While the benefits are great, software isn’t doing this by itself. In order for systems to sense, see patterns and make optimizations, they need a steady stream of data. Your city needs sensors and a communications infrastructure that can support the data they generate. As you read about the interesting things AI is doing for cities, remember it’s a smart cities backbone that ultimately makes them possible. — Kevin Ebi

Every week there are new stories about the incredible strides made by artificial intelligence — this week, AI detected 50 eye diseases as well as a doctor — but it’s also delivering real results for cities too. Here are three recent examples that are worth noting:

Reducing risk from complicated infrastructure projects
How do you accurately estimate the time and cost needed to do a complicated infrastructure project, like a long tunnel? The typical starting point is what you’ve done before, but when there are few (or no) similar projects to base yours on, what do you do then?

Bechtel, a North American Lead Partner of the Council, says this is an area where AI may be able to reduce risk. Just because a particular design may have never been done before, infrastructure projects do have a number of common elements. AI is great at finding them.

The company is building an AI-powered assistant that will be able to analyze unique projects and predict more likely outcomes based on the performance of unrelated projects. It’s starting with 10 years of disparate project data, but says it may one day be able to train AI to analyze 120 years worth of projects to find indicators in correlated areas that humans might miss.

Stretching overworked appraisers
Wake County, NC, has seen explosive growth. In less than two years, county appraisers have 400,000 parcels to assess, and 3,000 homes change hands every month. There’s no sign of it slowing down, either.

Oh, and the county is now re-assessing parcels every four years rather than its old eight-year cycle.

Wake County says there’s no way it could train all the appraisers it needs, even if it could find enough people to do the job, so it’s turned to an AI system from Council Global Lead Partner SAS.

The appraisal platform looks at each days sales and considers dozens of factors to forecast a value for each property in the system. It takes a detailed look at each property. For instance, whether the property has a vinyl or brick exterior can affect the price.

Humans still set the final values, but AI gives them a good starting point.

Spend less time at red lights
Nobody likes waiting at traffic lights, but AI could soon help shorten the delay. Some cities are already using real-time traffic data to adapt their signal timing, but AI could soon help reduce wait times for motorists even more.

The new approach lets intersections work together for the benefit of commuters across the city. As the light in one intersection turns green, it communicates the number of vehicles that have passed through to the next intersection, which determines its optimal timing for the best traffic flow.

Each intersection manages itself, but does so in harmony with the other nearby intersections. The intersections become more like a living ecosystem.

The system is based on Robotics Institute at Carnegie Mellon University and could reduce wait times at intersections by 40% and overall travel times by 25%.