Handcuffs Over AI: Solving Security Challenges With Law Enforcement

For the last 20 years, spending to solve cybersecurity goals has exploded to more than $100 billion annually. Security vendors tout advances in detection and protection that will provide the relief that has evaded the craft since its formation. The most recently heralded messiahs are machine learning (ML), security orchestration automation and response (SOAR), and artificial intelligence (AI).

For those of us who have been dumping budgets and hope into 20 years of broken promises, we find it very difficult to muster any degree of hope in the next round of techno-salvation. Five years ago, my suspicion turned to outright rebellious indignation and I started evaluating everything I believed to find a better path to sustainable cybersecurity operations (SecOps).

The first error that we made building SecOps is aligning its outcomes with those of information technology (IT). IT is an extension of manufacturing with success defined as the ability to create, ship, store, and transform data and services. Because much of SecOps craft knowledge came from IT, we built philosophies, semantics, and tactics to deliver the outcome of protecting and preserving IT services and data.

After a few decades of struggling to understand recurring and persistent failure in SecOps, the craft has begun to realize that SecOps has very little to do with IT. Instead, SecOps is a component of law enforcement and national security with outcomes demanding punitive actions to cybercriminals.

Another piece of cultural baggage we developed was separating digital life from real life. When users started logging on to the Internet in the 1990s, many found refuge in an anonymous, alternate reality. And we developed long-lasting perceptions that what happened in one reality did not affect what happened in the other.

In the meantime, healthcare, banking, shopping, work, and entertainment have become deeply intertwined with a digital counterpart. As the illusion of dichotomy has been removed, acts we once thought of as cybercrime are identified simply as crime.

The Power of Deterrence
With the age of hacker innocence gone, why does cybercrime prolifically persist? In 2016, the Department of Justice released a short paper entitled "Five Things About Deterrence." The primary finding of the essay is that "the certainty of being caught is a vastly more powerful deterrent than the punishment." The primary reason cybercriminals carry out crimes is a perception that they will not face punitive action.

The irony of the course of the craft of SecOps is we have done an extremely poor job in scaring criminals, while effectively terrifying one another. Fear, uncertainty, and doubt have long been the primary sales tools of cybersecurity vendors. To turn the tide of cybercrime, fear must be delivered to the criminals in the form of effective deterrence. Criminals must believe punitive action is probable when executing crimes.

Code of Silence
In cities like Chicago, the citizens of crime-ravaged communities fear the criminals more than they trust the police. The relationships between these communities and law enforcement are so strained that citizens do not provide evidence or testimony that will be used to successfully prosecute the criminals and guarantee deterrence.

The same outcome, born of different history, creates a lack of coordination between law enforcement and private organizations being targeted by cybercriminals. The logs and data in systems owned and maintained by these organizations contain critical information that would enable successful prosecution of cybercrime to become the norm, which would deliver deterrence.

Building SecOps on the incorrect outcomes of service and data availability have left the craft unprepared to align with law enforcement outcomes. The tools, workflows, and data provide little value to investigators and prosecutors. When an organization does report a crime to law enforcement, the responding agency must comb through a mess of disparate data locations and formats that is more complicated to process than a murder crime scene. It is an expensive and exhausting endeavor for law enforcement, which must be cautious with its resources.

In addition to the cost of processing a disorganized digital crime scene, private organizations that give unfettered access of their systems to law enforcement may expose themselves to unexpected risk. Records of a separate crime carried out by an employee may be discovered while investigating the reported crime. These discoveries can create legal and brand damage to the organization when prosecuted. The fear of these risks often hinder efforts to enable sustainable communications with law enforcement.

Moving Toward Sustainable Deterrence
To resolve the communication problems between organizations and law enforcement, a few shifts need to occur.

While operational risk management, governance, risk, and compliance are key components of cybersecurity efforts, the outcome of SecOps must move away from service availability through risk mitigation to crime deterrence through law enforcement coordination. With the fundamental outcome established, the way operations are conducted and evidence is handled will change. The result of most investigations will be a digital affidavit of evidence that law enforcement can use to successfully prosecute the criminals.

Additionally, anonymous tips to law enforcement are a key tactic to help law enforcement identify and hunt down the criminals. This type of information does not carry the same usefulness as an affidavit that can be used in court but does assist agents in detecting crimes and building cases.

To assist organizations in making the transition from IT-centric SecOps to law enforcement-focused SecOps, law enforcement agencies need to provide guidance on how they would prefer evidence to be collected, organized, and transmitted. Investments in community outreach and training will give agencies access to more evidence and allow them to process it more efficiently.