Navigating Autonomy: Unveiling Security Experts' Perspectives on Augmented Intelligence
summary
Grounded theory analysis of 27 in-depth interviews reveals a concrete decision framework: experts assign planning, strategy and uncertain decision‑making to humans and repetitive, high‑volume tasks (firewall config, network monitoring, report drafting) to AI, but treat penetration testing as mixed. Autonomy is judged across four autonomy levels—decision support, human approval, timed veto, and full automation—by weighing benefits, error impact, reversibility, urgency, and especially trust built via transparency and oversight.
AI-written summary · about
chapters (8)
- 0:00Augmented intelligence intro
- 2:19Research methodology
- 3:43AI capabilities and task sharing
- 6:14AI autonomy levels
- 7:18Factors in autonomy decisions
- 8:42Trust and verification
- 9:45Trust vs manual control
- 10:06Transparency importance
related talks
40mHuman-AI Collaboration Models for Better Decisions and Faster Workflows
20mSecurity in the Age of Agentic AI
21mSecurity in the Age of AI & Copilot
22mFireside Chat - AI Security Vulnerabilities
04mInterview
20mHuman above AI
15mAI Security Simplified: Turning Complexity into Cohesion
32mKeynote: Will the Real Autonomous Agent Please Stand Up