Rethinking Artificial Intelligence’s Role in Public Safety

Author: Bill Campbell, senior vice president of Global Public Safety for Hexagon's Safety, Infrastructure & Geospatial division

Over the last few years, Artificial Intelligence (AI) has been emerging as a key technology for public safety agencies wanting to build safer, smarter communities. AI helps bring data to the forefront of policing. In fact, 81 percent of public safety professionals believe that AI is important for their organizations.

While the benefits are significant, there is still lingering hesitancy about its use. For example, some policymakers, civil rights and technology organizations have raised ethical concerns around the use of certain AI capabilities for policing – and rightly so.

Much of the concern around AI hinges on the rise of predictive policing, which typically analyzes large sets of historic data to inform decision-making, patrolling and more.

While forecasting crime and other incidents might be well-intentioned, the data sets used to build these algorithms can have a range of biases baked into them. Therefore, police are rightly turning away from problematic uses. 

Yet the opportunities to improve public safety are clear. It’s time for public safety leaders to rethink the value of AI and move toward assistive uses in real-time situations, which can be a game-changer as emergencies unfold.

Time to Rethink the Role of AI in Public Safety

Predictive policing is all about using data from previous incidents to predict the future. The challenge is that historic data has inherent biases built into it. In other words, the past does not predict the future.

An emerging innovation called assistive AI can help police, fire and EMS respond to complex emergencies sooner through real-time pattern recognition.

For example, if there is a wreck involving a hazmat truck and a gas leak is detected, several calls could be coming in, but only one call includes info about the smell of gas. Another example is an active shooter situation where many calls are coming in with sparse and disconnected information for dispatchers.

Multiple dispatchers often also field separate calls about the same incident, making it difficult to create a coherent picture of what is actually happening. Or, during a staff shift change, incoming dispatchers may not have all the critical information from the previous shifts about a real-time event.

Assistive AI sifts through all of this operational data, such as call-taker and event notes and call transcriptions, to show 911 dispatchers patterns that can help determine if public safety events are linked – providing a unified picture of events as they unfold. This also removes the need for humans to review multiple and disparate data sources for connecting the dots around an event.

By detecting patterns and anomalies sooner, agencies can act faster and coordinate smarter to reduce the effects on communities, resources and staff. This is critical because the tide of data sources for public safety is increasing every day from IoT devices and other systems, which makes it difficult for analysts and operators to make sense of it all.

With an overwhelming amount of data from incidents, assets, cameras, units, alarms and more, assistive AI autonomously analyzes all information and issues proactive alerts based on observed trends and deviations.

And as more agencies consolidate 911 operations or collaborate more generally across jurisdictions, assistive AI can also help to improve information sharing for better decision-making.