AI’s Impact on Security in Miami-Dade and Pasco Counties

During a recent Sunshine Cyber Conference hosted by Cyber Florida in Tampa, panelists including Lawrence Embil, security manager for Miami-Dade County, and Larry Kraus, captain of the Pasco County Sheriff’s Office Research and Analysis Division, shared their perspectives on AI’s impact on risk and security as a whole.

Also weighing in on the discussion was JC Vega, a cybersecurity professional and retired U.S. Army colonel with decades of security leadership and cybersecurity operations experience.

To kick things off, panel moderator Michael Moore, a senior cybersecurity adviser at IBM, asked the panelists about potential security risks for the public, private and law enforcement sectors when using AI.

Embil shared a somewhat timely example impacting Miami-Dade County, stating, “We’ve all heard about the Francis Scott Key Bridge that was knocked down by the ship recently. Within 12 hours, we were already getting phishing emails to our employees with perfectly crafted malicious links, and we saw at least one person click the link before our system caught it and swept it up.”

As for how the example relates to AI, he explained how, due to an increase in automated scams targeting the county’s systems, county staff are using machine learning to reduce the processing of these events.

For example, Embil said, “I’ll give you a hypothetical number that’s going to be very close to the actual number; we’ll see like 30,000 events per month, and with machine learning, we can reduce those down to maybe 3,000. We’ll have another layer of analysis to go on and eventually get that down to around 15 to 30 events per month.”

Kraus said Pasco County is taking a multifaceted approach to AI.

“In essence, law enforcement collects a lot of data,” Kraus said. “We collect a lot of personally identifiable information, not to mention patterns and trends. We have had analysts that we have hired who have been able to go through the data, but not as fast as machine learning or AI does, right? So ultimately, law enforcement is using AI to be able to go across multiple data sets to basically federate searches to get faster at solving crimes.”

The flip side of AI being used in law enforcement is that the bad guys are using it, too.

“The cyber domain, for us, is a different domain. It moves fast, there’s a high reward and a low risk of getting caught,” Kraus said. “In some cases, we also are running into jurisdictional issues and things of that nature and identifying who the actual adversary is, right? We’re seeing a large number of threat vectors opening up in law enforcement.”

Another concern, similar to Miami-Dade County, is the “human element.”

“You want to be (talking) about phishing emails and things of that nature. The human element is killing us in law enforcement,” Kraus said. “You want to get a law enforcement officer to click on an email, give away free coffee, give away free 5.11 tactical pants — there’s a ton of ways.”

At the end of the day, however, protecting citizen data is a top concern.

“We’re starting to see that we’re a bigger target than we thought we were, and if we don’t do anything about it, it significantly puts our constituency, our public and our communities at risk because we are housing their data,” Kraus said.

More information about the conference can be found online.