AI/ML is now a key part of security

DevSecOps teams are becoming more broadly aware of security as a shared responsibility as they are incorporating security earlier in the software development lifecycle, or shifting left, according to the 2023 Global DevSecOps Report from GitLab.

This development is enabling development, security, and operations teams to work collaboratively instead of working in silos, as seen in previous years. 

Conducted by Savanta, a survey for the study was distributed via GitLab’s social media channels and email lists, and panel sampling. Responses were collected from 5,010 software professionals worldwide in March 2023.

Findings show that 71% of security professionals said that a quarter or more of all security vulnerabilities are being captured by developers, up from 53% of respondents in 2022.

Also, 38% of security professionals reported being part of a cross-functional team focused on security, up from 29% in 2022. 

Further, 85% of security respondents report that they have the same or less budget than 2022, highlighting an urgent need to do more with less.

Artificial Intelligence (AI) and machine learning (ML) have become critical components of DevSecOps workflows. Developers who use a DevSecOps platform were more likely to have implemented automation and AI/ML for testing than those who do not.

About two-thirds (65%) of developers said that they are using AI/ML in testing efforts or will be in the next three years.

More than three-fifths (62%) of developers using AI/ML use it to check code, up from 51% in 2022, and more than half (53%) of developers using AI/ML said they use bots for testing, up from 39% in 2022.

Developers and security professionals continue to report significant time spent on toolchain management, reducing time available to dedicate to critical tasks such as adherence to compliance regulations. 

Two in every three (66%) respondents reported wanting to consolidate their toolchains this year.

More than a quarter (27%) of security respondents reported that it is difficult to have consistent monitoring across disparate tools, and 26% of security respondents said it is difficult to draw cohesive insights across all integrated tools.

Despite ongoing demands for improved digital experiences within the public sector, respondents working within government entities globally noted slowed or stagnant software development. 

More than half of total government respondents said they are evaluating or purchasing a DevSecOps solution in one to three years.

Almost three-quarters (74%) of public sector respondents reported deploying software at the same rate or slower than they did in 2022, and 44% of public sector respondents reported using  or more tools for software development, including some who use more than 15 tools.

David DeSanto, chief product officer at GitLab, said that organisations globally are seeking out ways to do more with less. This means that efficiency and security cannot be mutually exclusive when identifying opportunities to remain competitive.

“DevSecOps tools and methodologies allow leadership to better secure and consolidate their disparate, fragmented toolchains and reduce spend, while also freeing up development teams to spend time on mission-critical responsibilities and innovative solutions,” said DeSanto.