How Good Is Your Phishing Training? NIST Launched a Tool to Figure That Out.

astel design/Shutterstock.com

Researchers are looking to pool more data from outside organizations to further refine it.

Researchers from the National Institute of Standards and Technology released a new mechanism to help information technology professionals better protect their organizations’ personnel from scams aimed at stealing their credentials.

Unveiled Thursday, the Phish Scale uses a NIST-crafted rating system to ultimately provide users with a firmer grasp of how their workforce may be vulnerable to phishing and other social engineering-type scams. Usually such scams involve calls or emails impersonating reputable sources to trick people into sharing personal or sensitive details such as passwords or financial information.

Those who made the Phish Scale now intend to team up with outside agencies and organizations to refine it further. 

“Initial phishing exercise data collection began back in 2012, so this really is the result of many years of NIST data and research,” Kristen Greene, a cognitive scientist on the multidisciplinary NIST team behind the scale, told Nextgov Friday. “Although many general phishing tactics from past years are still very active and applicable today, there are some disturbing trends in increasing phishing sophistication. In particular, targeted phishing attacks, i.e. spear-phishing, are on the rise.”

In the paper published detailing NIST’s development of the scale and how to use it, researchers confirm that “phishing in particular, and social engineering in general, are active threats across all industry verticals.” And such fraudulent attacks, as Greene mentioned, continue to evolve and surface—for instance, employees at social media giant Twitter recently fell for a spear-phishing attack that resulted in more than 100 profiles being compromised to induce people to fork out money in a major Bitcoin scam. 

Organizations often institute phishing training programs to prepare employees to vigilantly spot such scams, and chief information security officers and others who steer such programs generally focus on click rates, which present the frequency that users click on the fraudulent emails.

“Higher click rates are generally seen as bad because it means users failed to notice the email was a phish, while low click rates are often seen as good,” NIST’s release on the work notes, adding “however, numbers alone don’t tell the whole story.”

The Phish Scale provides users with a do-it-yourself method to determine why click rates implemented in their program are high or low. This, in turn, can help leadership puzzle out how to optimize and improve their phish-catching training initiatives. 

According to a video the agency released spotlighting the effort, the scale aims to classify emails based on how difficult or easy it is to detect possible phishing attacks. It implements two main components: observable “cues” or characteristics that might prompt the user to notice trickery; and what NIST deems “alignment of the emails’ context to the user,” which involves a full rating system. The agency notes that “emails with fewer cues and more relevant context” are the ones that make it most difficult to pinpoint as phishing.

“We hope that people move beyond focusing solely on click rates, and also have a corresponding understanding of the difficulty of their phishing exercises, to really help get at the impact training is having on the organization,” Shaneé Dawkins, a computer scientist also on the team that created the tool, told Nextgov. “The Phish Scale is a first step in that direction.”

Dawkins’ research partner Greene added that while “much phishing research” is conducted over short timeframes and inside laboratory settings, the data that underpins the Phish Scale was gathered over many years in a realistic workplace setting. The research paper that accompanies the work comprehensively details the teams’ collection methods.

“The phishing exercises represented an important variety of tactics and ranges of difficulty, some of which were very targeted spear-phish,” Greene said. “All mimicked real-world attacks and threats.”

All of the data used to date came directly from NIST, but going forward the researchers hope to broaden that pool and guarantee that the scale works across diverse operational environments.

“We encourage phishing training implementers to tailor their programs to be representative of current real-world threats facing their respective organizations,” Dawkins said. “We’re actively working to improve and simplify the Phish Scale, and are looking for agency partners to share data and help us validate and test the scale.”