Experts Advocate Real-Time Authorization Models for Autonomous AI Agents
Agentic AI systems, which can autonomously select tools and execute multi-step tasks, are expected to be integrated into 33% of enterprise software by 2028, up from less than 1% in 2024. This growing autonomy raises challenges for identity and access management, as traditional static permission models are inadequate. Security experts like Vatsal Gupta advocate for real-time, continuous authorization based on dynamic risk signals to ensure AI agents operate within appropriate boundaries as their context evolves.
First-hand measurement across 3 sources
We measured how 3 outlets covered this story. No outlet gave this story a measurable political slant — there is no left–right reading to report. Overall sentiment is neutral (52/100). Lens Score 30/100.
Outlets measured: oneindia, oneindia, oneindia. See how each one headlined and framed the same story in the source comparison below.
AI Analysis
Sentiment was consistent across outlets (52–52/100), indicating broadly factual reporting rather than editorialising.
Coverage timeline
oneindia broke this story on 3 Sept, 05:57 am. Other outlets followed.