The Real “Weakest Link” In Security Isn’t What You Think: Why We Should Rethink the Narrative That Humans Are What Make Us Less Secure
It’s an all-too familiar story: A company reports a data breach,and there’s an immediate blame game. Inevitably, we point the finger at humans — the person who responded to that phishing email ( a fake message that a bad actor uses to gain access to a broader set of data or a network) or who unknowingly clicked on ransomware “malvertising” (a fake ad that, when clicked, releases malware that locks digital files and demands a ransom to release the data).
Humans, we’re told, are the weak link of security. That was a key theme in the Verizon Data Breach Investigations Report released last week. After all, ransomware and phishing are effective because they’re able to so skillfully target human vulnerabilities.
Here’s the problem. Human vulnerabilities will always exist. This old way of thinking — that people are the problem, and we can somehow change entrenched human behavior — isn’t getting us anywhere. Even with improved training and education, given the sophistication of the attacks, human vulnerabilities will persist. So we need to rethink this paradigm: What if we started viewing human-computer interaction as a means toincrease security? How could we use what humans do best — critical thinking and contextualization- and combine it with what computers do best — automation and scale — to make us all safer?
We can start with a more “human-centric” approach to security — in other words, designing products and solutions with human strengths and vulnerabilities in mind. Here are three examples of ways that this approach could make us all more secure:
1) Alert fatigue — Monitoring systems with an overabundance of alerts aren’t just ineffective but lethal. With so many low priority alerts, users simply ignore the alerts or have little ability to differentiate between those high and low priority alerts. And given the vast amount of data, it’s impossible to respond to every single alert. For instance, at Target , the security team received and ignored alarms — in part because there were just so many. Many have pointed to this as human fallacy, but in reality it is a combination of human-computer interaction failure. With so many alerts, very few teams have the time or capabilities to sift through in depth every alert that is received. Even with the best judgment, systems with little ability to inform and prioritize alerts are simply ignored. In contrast, monitoring systems that integrate automation with human-driven domain expertise and prioritization could be a first step at more precise and relevant alerts, decreasing dwell time and expediting incident response.
2) Data exploration — Analyzing and protecting big data is getting more and more complicated as the amount of data that we generate increases, and as attackers begin to not only steal data, but to manipulate it, too. We need to create faster and more effective ways to explore the data required to analyze and detect intrusions, especially in the face of an industry-wide talent shortage. In short, there is too much data and too few people to analyze it, and this problem is only growing. So, how do we explore data faster and more efficiently? Cognitive methods aimed not just at supporting human hypotheses, but also proactively surfacing key insights will be an essential component for improved security. Machine learning and other forms of automation help scale these capabilities, and provide much faster insights than is possible through human analysis alone. For instance, in the commercial realm, cognitive computing helps answer customer and supplier questions, or in finance can identify optimal investment portfolios. These technologies help remove the arduous processes of data structuring and merging, but also provide optimized analytics so humans can devote their time to the important analysis, contextualization, and interpretation of the data required to detect and contain attacks. These tools do not replace the analyst, but provide greater, faster, and more scalable analytic capabilities to help prioritize and gain insights from data, greatly impacting detection and prevention of anomalous behavior. Automation and advanced data analytics also helps security teams optimize their resources, enabling greater detection capabilities of the seemingly infinite data with finite resources.
3) Mind the C-Suite Gap — It’s as high-stakes as communication struggles get: security teams often are unable to put their work and issues into language that CEOs can understand. When they can’t communicate effectively to company leaders, their warnings are disregarded, leading to devastating consequences. The C-suite increasingly bears the brunt of breaches — leading to turnover of CEOs and government leaders — but they may not grasp the complexities or resources required for security. Data visualization can bridge that gap. Think of it as the storytelling medium, conveying complex data in a consumable manner. Intuitive, interactive, and concise data visualization can express multifaceted concepts in a much more efficient manner than showing a presentation full of log data.
We hear a lot about changing human nature as the key to digital security. While education and training are essential, human behavior is nearly impossible to change and isn’t a silver bullet. Instead, let’s focus on building technologies that leverage the best parts of computers and humans working together. It could go a long way to address the increasingly complex challenges in the digital domain.
This post was previously published by New America.