As detailed in IT Pro, new research indicates that developers are failing to address security vulnerabilities 25% of the time, a situation increasingly complicated by the integration of artificial intelligence in coding workflows.
A study by SecureFlag analyzed over 800,000 attempts by developers to remediate software flaws. While 75% of these efforts were successful, the remaining 25% were incorrectly implemented. Researchers identified a significant knowledge gap, noting that developers often fail to recognize when a fix is inadequate. This issue is particularly concerning as the industry adopts AI tools to automate vulnerability remediation. Emilio Pinna, the chief technology officer at SecureFlag, warned that AI often generates plausible-looking code that may not actually resolve the underlying security issue. Without sufficient expertise, developers struggle to distinguish between effective solutions and ineffective, cosmetic rewrites. The problem is most acute in complex areas such as broken access control and software data integrity, where developers frequently fail to seek assistance. Industry experts emphasize that the primary value of security training is now enabling developers to critically evaluate AI-generated outputs rather than relying on them blindly. As AI usage grows, the inability to verify these automated fixes threatens to mask critical vulnerabilities rather than closing them.
Source: IT Pro
