top of page

Transforming the Frontlines: The Untapped Potential of AI in Public Safety

Updated: Feb 28

Recently, the National Institute for Science and Technology’s (NIST) Public Safety Communications Research (PSCR) lab and – my old stomping grounds – the First Responder Network Authority (FirstNet), hosted “5×5: The Public Safety Innovation Summit” in San Diego, California. For those who don’t know, FirstNet is the government authority that oversees AT&T’s FirstNet Network – a contract I led when I was at the Department of Commerce just a few short years ago resulting in the first and only network of its kind, dedicated to Public Safety nationwide.

For years, PSCR and FirstNet have partnered on several initiatives driving innovative technology into the hands of our first responders. In fact, 5×5 is the most recent evolution of the annual conference PSCR has been hosting since 2010, bringing together academia, industry, government, and the broader Public Safety community to accelerate adoption of new technology to improve response operations universally.

As usual, the conference had great attendance and hosted some very exciting tech. The showroom floor was filled with augmented and virtual reality hardware, deployable network tools, advance light detection and ranging (LiDAR) solutions, and more. There were several tracks attendees could follow with truly something for everyone:

  1. Location Services and Situational Awareness

  2. Mission Critical Services

  3. Network Services and Coverage

  4. Open Innovation and Commercialization

  5. Public Safety Operations

  6. User Interface / User Experience

  7. Secure Information Exchange

Embedded in each of these conversations were familiar topics for first responders; budget concerns, interoperability of solutions, spectrum management and availability for communications networks are common themes when this group gets together. And as much as I enjoyed seeing the fruits of our labor at FirstNet come to pass, I couldn’t help but feel there was something officially missing from the conference: Artificial Intelligence.

That isn’t to say the subject was avoided entirely by the technologists and Public Safety professionals in attendance. AI was very much the subject of many conversations, its potential impact on any technology is the current elephant in the room at many conferences like the 5×5. In my talks with attendees, there were sideline discussions about deploying intelligent large language models (LLM) in 9-1-1 centers as copilots to telecommunicators working to dispatch reponse resources. There was a discussion on the potential for AI to support forensic reviews via smart records management systems. And there was a lengthy discussion around regulation and governance of AI, and that AI’s capability far outweighs our legislative bodies’ ability to understand – and tolerate – its eventual ubiquity.

But before the regulatory walls come tumbling down on innovation as they appear to have done in the EU, the classically innovation-averse communities of practice like Public Safety are thankfully non-traditionally curious. It shouldn’t be surprising that technology adopted by our first responders should lag commercial innovation; things responders use in the field simply must work, as any one failure can be at the intersection between life and death. It wasn’t so long ago in 2014 when, as a program manager at FirstNet, I was often reminded that the technology that worked with certainty then could never be replaced by broadband as the networks themselves were unreliable, and therefore, not ‘mission critical’ . Challenging the status quo, AT&T built a more reliable network than any in the marketplace today, and it continues to improve as adoption of that network increases (nearly 5 million responders connect on it today). Adoption and use of technology consistently lead to improvements of that technology and subsequently, increased adoption.

Today, anyone can access a commercial instance of AI through large language models like Google’s Bard or ChatGPT (GPT 4.0 just became publicly available). Just as broadband was available to Public Safety before FirstNet, it was not a tool primarily adopted for operational use. In fact, broadband was one of a series of redundancies for many responders in the field if, for whatever reason, their land mobile radio (LMR) wasn’t working or there was a breakdown in dispatch. (I remember stories of a police chief carrying three phones, one from each carrier, to maximize availability in even the deadest coverage zones of the city).  Any technology on display at 5×5 could eventually find their way into the Public Safety response workflow, most likely as a redundancy at first, and perhaps an augmentation of function later.  The pedestal upon which responders place primary mission critical technology is small and crowded, mostly with solutions that fundamentally work but are far from innovative. Is there room for AI in mission critical services? Maybe not yet, but eventually. But I’d bet good money that there will never be an AI-enable radio handset over an LMR network.

Stepping back for a minute… AI may not immediately manifest as an intelligent voice in a responder’s headset or helmet as J.A.R.V.I.S does in Iron Man. (And yes, that was an acronym. Just A Rather Very Intelligent System.) It may appear in a more subtle manner, helping to shape criminal cases by culling data collected during an investigation into a prosecutorial narrative. It may be an in-vehicle voice assistant, providing real time tactical directions to a cruiser in pursuit of a suspect using vast access to city internet-of-things data, like a GPS on steroids. It may also be a passive listener to 9-1-1 calls providing call takers and dispatchers with key insights trigger by words used, voice sentiment, or even caller sidelink data to paint a more precise picture for responders as they arrive on scene.

None of these scenarios are far out of reach. In fact, as I walked the show hall for the 5×5 conference, many of the technologies on display could one day provide either the network to transit the data necessary to feed an AI ecosystem for Public Safety or be the beginning data point to help inform responders on how they should best execute their job at any given incident. As with most institutions, it’s about being ready to ingest the capability into your workflow and manage the change to your organization, however large or small. Some organizations will take longer than others to adjust, and those that do will more than likely get it right in ways others who are quick to adopt may not.

For Public Safety, the first step to innovating beyond LMR and their traditional approach to response was the government’s investment in the FirstNet program. While that may not seem like a big leap, remember, it took almost 10 years for Congress to put legislation in place to make FirstNet actually happen – one of the last remaining requirements of the 9-1-1 Commission Report was a dedicated broadband network for Public Safety. That delay could also explain the delay in the kinds of technology we’re seeing Public Safety invest in. In many ways, for the technologists at the 5×5 conference, the first step to getting access to users was also FirstNet.

For the open-source AI community kicking the tires on LLMs and what they mean for any number of industries, FirstNet may also be their first step in connecting to Public Safety. Time will tell. For now, like the conversations at the 5×5 conference that surrounded the main show floor, AI is just a bit farther out of sight, ready to be accessed and deployed for Public Safety when the time is right.

Looking to make a change to how you manage your organization? Are you curious about preparing for AI deployments to support your workforce? Reach out directly if you have any questions or would like a better understanding about how Itero Group can help you on your AI journey.


Couldn’t Load Comments
It looks like there was a technical problem. Try reconnecting or refreshing the page.
bottom of page