Loi Liang Yang Says AI Agents Will Reshape Cybersecurity Work

Summary

Loi Liang Yang argues that AI agents could automate security-log triage, identity-permission reviews and other workflows, urging cybersecurity professionals to retain core skills while learning to operationalize agents.

Key points

• Yang says reviewing firewall, antivirus, operating-system and cloud logs can take security analysts hours or days to distinguish false positives from incidents.
• He describes agentic AI as able to take actions on a user's behalf, rather than only return answers.
• Examples he gives for security tasks include reviewing identity permissions, remediating website misconfigurations, correlating logs and applying patches or virtual patches with approval.
• He warns that manual querying and evidence gathering may be automated, while identifying agent deployment, management and workflow automation as opportunities.
• He recommends learning to operationalize AI agents while retaining foundational cybersecurity skills.
• He suggests MCP servers could connect agents to security tools and support investigation, triage, containment and remediation workflows.

People