SovereignPA-Bench: Evaluating User-Owned Personal Agents under Evolving Intent, Platform Mediation, and Consent Constraints

arXiv:2607.05363 · cs.AI · Submitted 2026-07-06 · Read on arXiv

cs.AI

Submitted: 2026-07-06

Updated: 2026-10-01

Code: https://github.com/microsoft/social-reasoning-bench

Terminology

• Benchmark — 3× in this paper · explained in 2 episodes
• Alignment — 1× in this paper · explained in 1 episode
• Personalization — 1× in this paper · explained in 2 episodes
• Scaffolding — 1× in this paper · explained in 3 episodes

Sources

• Firewalls to Secure Dynamic LLM Agentic Networks
• AI loyalty: A New Paradigm for Aligning Stakeholder Interests
• What Is Your AI Agent Buying? Evaluation, Biases, Model Dependence, & Emerging Implications for Agentic E-Commerce
• AirGapAgent: Protecting Privacy-Conscious Conversational Agents
• Magentic Marketplace: An Open-Source Environment for Studying Agentic Markets
• $\tau^2$-Bench: Evaluating Conversational Agents in a Dual-Control Environment
• Designing Fiduciary Artificial Intelligence
• STALE: Can LLM Agents Know When Their Memories Are No Longer Valid?
• AppWorld-UL: Benchmarking Diverse Agent-User Interactions for Tool-Use
• KnowU-Bench: Towards Interactive, Proactive, and Personalized Mobile Agent Evaluation
• CI-Bench: Benchmarking Contextual Integrity of AI Assistants on Synthetic Data
• DECEPTICON: How Dark Patterns Manipulate Web Agents
• AgentDojo: A Dynamic Environment to Evaluate Prompt Injection Attacks and Defenses for LLM Agents
• Value of Information: A Framework for Human-Agent Communication
• Investigating the Impact of Dark Patterns on LLM-Based Web Agents
• WASP: Benchmarking Web Agent Security Against Prompt Injection Attacks
• The Ethics of Advanced AI Assistants
• Operationalizing Contextual Integrity in Privacy-Conscious Assistants
• Capable but Careless: Do Computer-Use Agents Follow Contextual Integrity?
• ConVerse: Benchmarking Contextual Safety in Agent-to-Agent Conversations

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems