Pierluigi Paganini November 06, 2018

The National Institute of Standards and Technology (NIST) is planning to use Artificial Intelligence to assign the CVSS scores to reported vulnerabilities.

The Common Vulnerabilities and Exposures (CVE) system provides a reference-method for publicly known information-security vulnerabilities and exposures.

A Common Vulnerability Scoring System (CVSS) score between 0.0 and 10.0 that is assigned to each flaw according to its severity. The numerical score can then be associated with a qualitative representation (such as low, medium, high, and critical) to help organizations properly assess and prioritize the issue.

The NIST will use IBM’s Watson to automatically evaluate the level of severity for each reported vulnerability and assign a proper severity score.

The CVSS score depends on some factors such as the complexity of the attack for the exploitation of the flaw, the effect of the attack on confidentiality, integrity, and availability of the target system, the size of the audience impacted, whether the attack requires the user’s interaction, and whether the flaw could be exploited remotely.

Currently, the CVSS scores are assigned by experts at the NIST, but the organizations believe that the introduction of the AI could speed up the process and increase its efficiency.

According to Matthew Scholl, chief of the National Institute of Standards and Technology’s computer security division, the AI will replace human analysts by October 2019.

An analyst takes 5 up to 10 minutes to assign a score to a simple vulnerability, but the time could be far longer for more complex issues. Scholl pointed out that the number of vulnerabilities publicly disclosed is increasing with each passing year.

“Earlier this year, NIST launched a pilot program using IBM’s Watson artificial intelligence system to pore through hundreds of thousands of historical CVSS scores from the institute’s human analysts, Scholl said.” reported NextGov.

Watson used the data to build its experience and use it to assign scores to new vulnerabilities.

“We started it just to get familiar with AI, so we could get our hands on it, learn about it, kind of put it in a lab and experiment,” Scholl said. “As we were doing it with this dataset we said: ‘Hey, this seems to be putting out results the same as our analysts are putting out.’”

The scores assigned by Watson were similar to the ones provided by analysts for not complex flaws of for vulnerabilities with many similarities to previously reported ones.

“The Watson system is great at assigning scores for vulnerabilities where there’s a long paper trail of human-assigned scores for highly similar vulnerabilities. In those cases, the Watson score will be within the small range of variance between what two different human analysts would assign, say 7.2 versus 7.3 on a 10-point scale, Scholl said.” continues the NextGov.

“When the vulnerability is new and complex or highly novel, like the Specter vulnerability discovered in 2017, Watson fares far worse, Scholl said. In those cases, a human analyst will take over.” 

IBM Watson also releases a confidence percentage for each CVSS score, when the percentage is under a specific threshold it is requested the analysis of a human analyst.

They are also looking into using the technology in other NIST areas.

Pierluigi Paganini

(Security Affairs – CVSS scores, NIST)

[adrotate banner=”5″]

[adrotate banner=”13″]