BeyondTrust Finds 75% of Cyberattacks Stem from Identity and Privilege Gaps — BeyondTrust Corporation

Event summary

BeyondTrust's Phantom Labs® Research Index analyzed 400+ offensive research projects, finding 75% of cyberattacks traced back to identity or privilege issues.
Top root causes included credential exposure (18%), identity relationship exposure (11%), and excessive standing privileges (11%).
AI agent security was a major focus, with 50% of research projects centered on AI platforms and agents.
Coordinated vulnerability disclosures were made in OpenAI Codex and AWS Bedrock AgentCore.

The big picture

BeyondTrust's findings highlight a critical shift in cybersecurity, where attackers increasingly exploit trusted relationships between identities rather than isolated vulnerabilities. This trend is particularly pronounced in AI-driven environments, where machine and AI agent identities often operate with minimal oversight. The research underscores the need for comprehensive privilege-centric security frameworks across cloud, SaaS, and on-premises ecosystems.

What we're watching

AI Security Risks: How the rapid deployment of AI agents as enterprise identities will impact security frameworks and compliance.
Privilege Management: Whether organizations can effectively reduce standing privileges without disrupting operational workflows.
Cloud Ecosystem Vulnerabilities: The pace at which major cloud providers like AWS and Microsoft address identity-related security gaps.