The role of AI in federal facility safety

The U.S. federal government is tasked with securing more than 113,000 buildings and other facilities nationwide, an enormous responsibility with seemingly endless shortcomings.

Technology is often at the forefront of these efforts, and these systems are quickly becoming outdated and ineffective. According to one Federal News Network analysis, existing security systems and cameras are in various states of disrepair, requiring a strategic plan “to repair, replace and even install security cameras and alarm systems as part of a nationwide assessment.”

In an era of heightened risk of attacks on government properties and the increase in mass violence incidents, this isn’t just a passive directive. It is an urgent priority, especially with the 2024 election cycle just around the corner.

Unfortunately, government agencies and organizations are too often slow to adopt new technologies, as a lack of knowledge or fidelity to existing guidelines stifles innovation and diminishes potential outcomes.

We can do better.

It’s time for a digital transformation in our federal building security apparatus. In particular, as the age of artificial intelligence comes more fully to fruition, we have the opportunity to leverage this technology to prioritize efficiency without compromising efficacy.

Elevating security without compromising convenience

AI security solutions can automate defense mechanisms, increase the speed and efficiency of threat detection, and enhance manual intervention.

Many of these technologies allow buildings to maintain a natural traffic flow, ensuring safety without causing unnecessary congestion, anxiety or inconvenience for those passing through them. With more than 1.4 million employees and visitors entering federal buildings daily, elevating security without compromising convenience is critical.

However, existing solutions, like metal detectors or manual pat downs, require people to empty their pockets and undergo thorough checks and clog corridors, creating discomfort, annoyance and a sense of being treated suspiciously. More importantly, those solutions give a false sense of security depending on metal detectors being 100% accurate and assume zero social engineering – the use of psychological manipulation to trick someone into making security mistakes – on behalf of the threat actor. These technologies are not 100% perfect and threat actors will always look for ways to subvert humans.

Instead, the goal is to implement security measures that effectively identify potential threats (like weapons) but are discreet enough not to disrupt the normal flow of activities. For instance, advanced AI-driven systems accurately detect guns, knives and other threats, reducing confrontations and improving entry speed. They also take the social engineering equation out of the loop as divestment is no longer the key to identifying threats.

Existing manual processes are straining resources as agencies look to level up their security footprint to account for emerging threat patterns. When AI-powered security solutions protect federal buildings, agencies maintain a defensive posture even when budget constraints (or more frequent budget showdowns) put building security at risk.

Of course, these technologies alone won’t make federal buildings more secure. When they work in tandem with security personnel and established processes to augment and enhance a team’s capabilities, they can do more even as potential threats continue to escalate.

Ethical, secure AI is critical

If AI will elevate security, we must ensure it does so without bias.

Specifically, the development of AI systems must be underpinned by robust and diverse training data to avoid any unintended prejudicial outcomes. For example, training needs to determine threats, not classify gender or ethnicity. This means training models with all kinds of people while ensuring the threats stay the same and compiling training data from different locations/regions.

Similarly, as the technology evolves, it’s crucial to ensure that AI operates without bias on weapons. More specifically, weapons are always evolving, and those threats must be continuously added to the data that training models use.

It’s also important to consider and prioritize how threats vary by region. Benign objects must also be introduced by what people in a given region/location usually wear.

Simply put, safe, effective AI can elevate security, but it requires intentional and strategic training to maximize its greatest assets and minimize potential drawbacks.

New security solutions protect against a novel threat landscape

As we navigate a landscape of evolving threats and heightened security needs, integrating AI-driven solutions in federal facilities is not just a forward-thinking move but a necessary evolution.

It’s time to move beyond the solutions to past problems, utilizing the latest innovations and operational efficiencies to make buildings (and our people) safer than ever.

This shift towards AI-enabled security systems represents a proactive step in safeguarding our nation’s federal facilities, ensuring that they remain accessible and secure without compromise.

With an election cycle on the horizon, now is the time to make these meaningful upgrades to our security footprint.

Joshua Douglas, senior vice president for product and engineering at Xtract One Technologies, is a recognized expert in the fields of behavioral analytics, cybersecurity and physical security working within schools, healthcare, government, critical infrastructure and commercial business.

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.