AUTHORITY & TRUST

Dr. James Joshi

Professor; Founder & Director, LERSAIS

School of Computing and Information, University of Pittsburgh

James Joshi’s research focuses on access control, trust management, and privacy in distributed systems and AI/ML — including attribute-based access control, insider-threat detection, and security in IoT, edge, and cloud environments. He founded LERSAIS, a designated Center of Academic Excellence in Cybersecurity Education and Research.

IEEE Fellow (2023, “for contribution to access control and privacy”)

PhD, Purdue University

NSF CAREER Award, 2006

Distinguished Member, ACM

the partnership

Authority that holds across every delegation.

Agent systems need more than permissions. They need authority that can be established, delegated, checked, and continuously re-evaluated as identities, contexts, and risks evolve.

James Joshi’s work on access control, trust management, and insider-threat resilience makes that authority layer operational — from the first credential to every consequential agent-to-agent handoff.

the research

Authority and trust, verified continuously.

Three authority-and-trust directions translate Dr. Joshi’s access-control and security research into accountable agent operations.

DIRECTION 01

Attribute-Based Identity & Access for Agent Fleets

Model identity and authorization as rich, contextual attributes that can express what an agent may do, for whom, and under which conditions.

Identity architecture · access control · policy context

DIRECTION 02

Secure Agent-to-Agent Delegation

Make delegation explicit, revocable, and auditable so authority cannot quietly expand as work moves between agents and organizations.

Delegation chains · signed authority · cross-domain trust

DIRECTION 03

Continuous Trust Evaluation Under Insider-Threat Conditions

Continuously assess evolving agent behavior and context so insider risk is detected before trust assumptions turn into exposure.

Trust evaluation · anomaly context · insider resilience

supporting research

Authority, privacy, and trust under real operating conditions.

Advanced Access Control & Trust Management

Develop access-control and trust-management models that keep authority precise across distributed, AI/ML, IoT, edge, and cloud environments.

Policy models · identity context · accountable access

Privacy-Preserving AI/ML & Insider-Threat Resilience

Protect sensitive decision systems while detecting and mitigating risks that emerge from trusted but compromised insiders.

Privacy-preserving AI · insider threat · continuous assurance

← Back to research portfolio