The latest Hacker-Powered Security Report from HackerOne reveals both optimism and apprehension within the security community regarding artificial intelligence, according to TechRepublic.
Based on surveys of security experts, members, and customers as well as insights from platform data, the greatest concerns were shown to be over leaked AI training data at 35%, followed by unauthorized use at 33% and the vulnerability of AI models to external manipulation at 32%. Nearly half of security leaders see AI as their organization’s top security risk, with data integrity emerging as a primary issue.
The report further noted that only 38% of HackerOne customers feel equipped to defend against AI-specific threats, while a separate SANS Institute study found a majority of security professionals expect an “arms race” between defenders and attackers using AI. Within security programs, AI’s primary applications include anomaly and malware detection, although respondents also cited challenges such as frequent false positives and limited effectiveness against novel threats. In light of these findings, HackerOne advises organizations to establish robust AI governance frameworks, incorporate regular validation processes, and offer employee training on AI security practices to address potential vulnerabilities.
