Intelligent automation plugs gaps in inmate threat intelligence

21/07/2021

Improving Upon Inmate Intelligence

By Johnmichael O’Hare

County sheriff’s departments carry the responsibility for running most of the nation’s jails, but most lack a critical tool for managing them effectively: timely and accurate inmate threat intelligence.

The National Sheriffs’ Association reports that more than 85% of jails in the U.S. fall under the sheriffs’ purview. Those facilities rank among the largest sources of litigation for sheriffs, according to the association. When inmates die due to homicide or fatal contraband drug overdoses, lawsuits often follow.

Inmate threat intelligence, however, can serve as an early warning system that helps sheriffs keep inmates safe from each other, while also reducing the potential for drug smuggling. Unfortunately, sheriffs’ agencies typically find it difficult to collect and analyze information of inmates that could foster heightened situational awareness.

Obstacles in the path of intelligence gathering

Disparate systems are one obstacle. A sheriff’s department might have an intelligence database and an inmate records system, which it may try to cross-reference with law enforcement records. It’s a cumbersome task to pull that information together to glean insight on inmates and their affiliations.

Information sharing among county jurisdictions is similarly difficult since unified intelligence-gathering and records management systems don’t exist at that level. Whatever information exchange that actually takes place tends to happen through phone calls, email messages, and monthly or quarterly cross-agency intelligence-sharing meetings. These traditional methods provide some value. Periodic intelligence meetings may shed light on emerging trends, providing a heads-up regarding a group of inmates currently at odds. But these high-level discussions rarely provide actionable information on certain inmates and their internal and external affiliations.

Another challenge: Sheriff’s agencies rely almost entirely on manual approaches for collecting inmate threat intelligence. Having a deputy check a newly arrived inmate’s social network presence provides a quick way to check for relationships that could indicate the potential for trouble. Sheriff’s personnel, given the constant influx of inmates, will be hard-pressed to find relevant information on arrivals as they manually scan the wide array of web sources. They may get lucky occasionally, but they will likely miss critical information. Manual methods can’t scale to deal with the number of inmates and their digital footprints. Manual searching could help a sheriff’s intelligence team collect useful information on, at best, a half dozen inmates – but that will barely scratch the surface of a jail population rife with inmate subcultures and ecosystems. It’s impossible to keep up using strictly manual methods – there are simply too many platforms to check.

While mainstream web data exist on the “surface web,” other platforms, forums, and websites reside on the dark web. This online tier, accessible only through a specialized router for anonymous browsing, has become a common destination for threat actors, including home-grown violent extremists. Some inmates could well leave an information trail on dark websites, which requires tools and expertise to probe effectively. The dark web could become a knowledge gap, depending on a jail’s inmate mix. But playing catchup using manual mining will prove ineffective, due to the number and variety of sites.

Finally, resource constraints and organizational culture can impede inmate intelligence gathering. Few sheriff’s agencies will assign more than a couple of people to conduct online searches. And that small cadre will likely face pushback if they don’t immediately produce results. The situation is analogous to physical surveillance, where an 80-hour stakeout may not turn up anything useful. Both types of observation require consistent application of effort – and a good deal of patience. But the perception that the watchers are “sitting around and doing nothing” typically cuts short such endeavors.

The resource issue exacerbates the dependence on labor-intensive investigative strategies. A few deputies conducting manual open web or dark web searches are in for a long, and potentially fruitless, struggle. Added together, isolated systems, manual methods, and limited resources result in an inmate threat intelligence process that is complicated, time-consuming, and ineffective.

Threats and their implications

Inadequate intelligence-gathering increases risk exposure in an environment where threats abound. New inmates admitted to the jail will have external and, possibly, internal connections. Failure to identify those relationships can have disastrous consequences. Violence is likely to erupt if an inmate affiliated with a particular security threat group is placed in a cell, or a section of a jail, with representatives of rival groups. Corrections officers, swept up in the ensuing chaos, are also put at risk. Proper placement is paramount for everyone’s safety.

A jail that doesn’t know its inmates can rapidly destabilize. But violence among inmates isn’t the only threat. When corrections officials can’t fully determine inmate affiliations, prisoners can persist in managing their business dealings with internal and external networks. In addition, an inmate associated with a corrupt guard or another staff member could become a conduit for contraband. Illicit drugs, such as fentanyl, entering a prison can spark a wave of overdosing. Prisoner violence, inter-group conflict, and increased contraband smuggling – and the looming potential for loss of life – all become more likely amid poor inmate intelligence. Other repercussions could include negative media coverage and state investigations or sanctions.

The stakes are clearly high, but a county jail’s situational awareness is often low.

Boosting intelligence for threat mitigation

So, how can sheriff’s departments mitigate threats and know their inmates better? Extracting data and insights from the open web and other web sources can put officials on a much better footing. But intelligent automation, rather than complete dependence on manual methods, is the key to making effective intelligence gathering a reality. This approach dramatically improves an agency’s ability to gather actionable intelligence on inmates and identify potential problems before they become tomorrow’s headlines.

Indeed, intelligent automation is critical for running online investigations at scale. Automation can provide continuous surveillance, relieving human investigators from the grind of spending hours behind a keyboard conducting manual searches. Artificial intelligence (AI), when added to automation, becomes a force multiplier. Automated searches using natural language processing, a branch of AI, lets investigators conduct complex keyword searches using multiple parameters. A sheriff’s department running a jail, for example, could build a custom query using gang-indigenous terms. Automation means that AI-enabled search can simultaneously scan multiple open web data sources as well as hard-to-penetrate dark websites.

Automated alerts can then bring relevant information to the attention of sheriff’s officials. Departments can deploy their limited resources to act on promising leads, rather than spend hours searching the immensity of the internet for a nugget of intelligence. AI and the related field of machine learning (ML) also help with scaling intelligence gathering and analyses, addressing the millions of digital footprints that surface every day. ML algorithms, for instance, can be trained to filter information to focus on just the types of data investigators wish to target. Searchers can find their needles without manually sorting through the haystack.

The image recognition aspect of AI, meanwhile, offers a visual dimension to searching web data, which increasingly involve images and videos. Audio-visual inquires can yield important intelligence for jail managers. Gangs, which often maintain relationships with internal security threat groups, have been known to drop hints regarding allies and outgroups in homemade music videos.

Benefits of intelligent automation

These components of intelligent automation provide jail managers much greater visibility. The ability to explore social connections – revealed through web intelligence (WEBINT) – provides insight into organizational structures within the jail and their extensions into the outside world. Importantly, intelligence automation taps sources of inmate threat intelligence traditionally beyond the reach of agencies. Sheriffs’ departments may take the jail’s temperature based strictly on what they observe in the facility. Visual analysis is probably limited to the jail managers’ knowledge of inmate tattoos featuring gang insignia. In any event, they could be overlooking important pieces of intelligence.

For sheriff’s departments, some of the benefits of AI-based web intelligence and automation include:

  • Reducing the potential for violent confrontations among inmates. Customized searches spanning the multitude of SOCMINT platforms shed light on an incoming inmate’s friends and foes. As for the latter, gang members aren’t shy about naming people they don’t like in their posts. Learning about those relations lets jails plan for safe placement.
  • Uncovering potential threats against jail staff. SOCMINT can reveal relationships across two or three degrees of separation. An automated check on social connections may find a new inmate has a problem with an incarcerated rival whose sibling is a nurse at the jail.
  • Improving knowledge of jail networks. Assessing digital footprints may help officials uncover previously overlooked nodes in security threat networks.
  • Achieving continuous situational awareness. Automated systems don’t go home at the end of a shift. Alerts keep officials apprised of potential problems, without assigning staff to hours of social-media duty.
  • Keeping tabs on the dark web. The dark web, while probably not the preferred haunt of most inmates, has become attractive to extremists who have been blocked from mainstream social networks. Information obtained from the dark web could reveal trends worth keeping an eye on.
  • Providing a single pane of glass. The task of integrating standalone intelligence and records systems into a unified system is very likely to be cost-prohibitive or technically unfeasible. Or both. Intelligent automation, however, pulls together information from each of those systems, providing a single-pane-of-glass view of data. While it may prove impossible to unify the systems, at least it’s possible to unify the view.

Exploring a new world of intelligence

A sheriff’s department responsible for managing a jail needs a comprehensive view of inmates. Traditional inmate threat intelligence gathering methods, while still useful, can’t provide the complete picture. New arrivals at county jails leave their mark on social networks, forums, and dark websites – much as they do in the physical world of cities, towns, and neighborhoods.

It makes sense to search online for inmate information. But the vastness of cyberspace, coupled with agencies’ resource limitations and manual methods, makes it difficult to effectively comb the web for useful data. Intelligent automation, however, offers the ability to tackle the web’s immense storehouse of information, conduct deep analysis and expose the social connections jail managers need to know about.

There’s a whole world of critical information out there.

The combination of automation and AI helps organizations make use of it.

Source: National Sheriffs’ Association Page 36-37

Share:

Skip to content