AI Security Threats Outpace Organizational Preparedness, O'Reilly Survey Finds

  • 55% of security professionals report encountering AI-generated phishing attacks, yet 52% say their work rarely or never involves AI-generated threats.
  • Only 44% of organizations have a comprehensive AI security policy in place, despite 49% of respondents citing overreliance on AI tools as a top concern.
  • Specialized AI security tools remain underutilized, with only 38% using AI for vulnerability scanning and 27% for threat detection.
  • 7% of respondents feel no skills gap, while 34% want to use AI tools more effectively and 24% want to learn how to secure systems using AI.

The survey highlights a critical inflection point in cybersecurity, where AI-generated attacks are becoming commonplace but organizational preparedness lags. This gap is exacerbated by uneven funding, underutilized specialized tools, and a significant skills gap among security professionals. The findings underscore the need for faster adoption of AI-specific policies, tools, and training to keep pace with evolving threats.

Policy Enforcement
Whether organizations with comprehensive AI policies can sustain lower rates of shadow AI use and better funding for AI security.
Tool Adoption
The pace at which specialized AI security tools will be adopted, given their reported effectiveness in improving alert triage.
Emerging Threats
How the underestimation of threats like prompt injection and data poisoning will impact organizations as they adopt locally specialized AI models.