Cover Image

Safety is Non-Compositional: A Formal Framework for Capability-Based AI Systems

Fuente: arXiv

Saved in:

Bibliographic Details
Main Author:	Spera, Cosimo
Format:	Preprint
Published:	2026
Subjects:	Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

From Workflow Automation to Capability Closure: A Formal Framework for Safe and Revenue-Aware Customer Service AI
by: Spera, Cosimo
Published: (2026)

Capability Safety as Datalog: A Foundational Equivalence
by: Spera, Cosimo
Published: (2026)

Safety as Computation: Certified Answer Reuse via Capability Closure in Task-Oriented Dialogue
by: Spera, Cosimo
Published: (2026)

Reversing the Paradigm: Building AI-First Systems with Human Guidance
by: Spera, Cosimo, et al.
Published: (2025)

Redefining CX with Agentic AI: Minerva CQ Case Study
by: Agrawal, Garima, et al.
Published: (2025)

Beyond-RAG: Question Identification and Answer Generation in Real-Time Conversations
by: Agrawal, Garima, et al.
Published: (2024)

BlueGlass: A Framework for Composite AI Safety
by: Nandigramwar, Harshal, et al.
Published: (2025)

Formalizing the Safety, Security, and Functional Properties of Agentic AI Systems
by: Allegrini, Edoardo, et al.
Published: (2025)

Safe Bilevel Delegation (SBD): A Formal Framework for Runtime Delegation Safety in Multi-Agent Systems
by: Sun, Yuan
Published: (2026)

A Conceptual Framework for AI Capability Evaluations
by: Carro, María Victoria, et al.
Published: (2025)

With Great Capabilities Come Great Responsibilities: Introducing the Agentic Risk & Capability Framework for Governing Agentic AI Systems
by: Khoo, Shaun, et al.
Published: (2025)

Evaluating Human-AI Safety: A Framework for Measuring Harmful Capability Uplift
by: Vaccaro, Michelle, et al.
Published: (2026)

HySafe-AI: Hybrid Safety Architectural Analysis Framework for AI Systems: A Case Study
by: Pitale, Mandar, et al.
Published: (2025)

Using Formal Models, Safety Shields and Certified Control to Validate AI-Based Train Systems
by: Gruteser, Jan, et al.
Published: (2024)

A Framework for Evaluating Emerging Cyberattack Capabilities of AI
by: Rodriguez, Mikel, et al.
Published: (2025)

SENTINEL: A Multi-Level Formal Framework for Safety Evaluation of Foundation Model-based Embodied Agents
by: Zhan, Simon Sinong, et al.
Published: (2025)

The AI Pyramid A Conceptual Framework for Workforce Capability in the Age of AI
by: Khatri, Alok, et al.
Published: (2026)

DeepKnown-Guard: A Proprietary Model-Based Safety Response Framework for AI Agents
by: Li, Qi, et al.
Published: (2025)

Let's Reason Formally: Natural-Formal Hybrid Reasoning Enhances LLM's Math Capability
by: Wang, Ruida, et al.
Published: (2025)

Defining Operational Conditions for Safety-Critical AI-Based Systems from Data
by: Christensen, Johann Maximilian, et al.
Published: (2026)

OpenAgentSafety: A Comprehensive Framework for Evaluating Real-World AI Agent Safety
by: Vijayvargiya, Sanidhya, et al.
Published: (2025)

AI Safety is Stuck in Technical Terms -- A System Safety Response to the International AI Safety Report
by: Dobbe, Roel
Published: (2025)

Semantic Entanglement in Vector-Based Retrieval: A Formal Framework and Context-Conditioned Disentanglement Pipeline for Agentic RAG Systems
by: Loghmani, Nick
Published: (2026)

AI Safety as Control of Irreversibility: A Systems Framework for Decision-Energy and Sovereignty Boundaries
by: Shu, Wesley, et al.
Published: (2026)

Domain-Agnostic Scalable AI Safety Ensuring Framework
by: Kim, Beomjun, et al.
Published: (2025)

X-RAY: Mapping LLM Reasoning Capability via Formalized and Calibrated Probes
by: Tianxi, Gao, et al.
Published: (2026)

SafeEmbodAI: a Safety Framework for Mobile Robots in Embodied AI Systems
by: Zhang, Wenxiao, et al.
Published: (2024)

PACEbench: A Framework for Evaluating Practical AI Cyber-Exploitation Capabilities
by: Liu, Zicheng, et al.
Published: (2025)

A Formal Security Framework for MCP-Based AI Agents: Threat Taxonomy, Verification Models, and Defense Mechanisms
by: Acharya, Nirajan, et al.
Published: (2026)

FormalGeo: An Extensible Formalized Framework for Olympiad Geometric Problem Solving
by: Zhang, Xiaokai, et al.
Published: (2023)

Safety Cases: How to Justify the Safety of Advanced AI Systems
by: Clymer, Joshua, et al.
Published: (2024)

Emerging Practices in Frontier AI Safety Frameworks
by: Buhl, Marie Davidsen, et al.
Published: (2025)

A Defect Classification Framework for AI-Based Software Systems (AI-ODC)
by: Alannsary, Mohammed O.
Published: (2025)

Quantifying Multimodal Capabilities: Formal Generalization Guarantees in Pairwise Metric Learning
by: Zhou, Richeng, et al.
Published: (2026)

FVEval: Understanding Language Model Capabilities in Formal Verification of Digital Hardware
by: Kang, Minwoo, et al.
Published: (2024)

Fighting AI with AI: Leveraging Foundation Models for Assuring AI-Enabled Safety-Critical Systems
by: Mavridou, Anastasia, et al.
Published: (2025)

The Institutional Scaling Law: Non-Monotonic Fitness, Capability-Trust Divergence, and Symbiogenetic Scaling in Generative AI
by: Baciak, Mark, et al.
Published: (2026)

A Dynamical Systems Framework for Reinforcement Learning Safety and Robustness Verification
by: Nasir, Ahmed, et al.
Published: (2025)

AI-Educational Development Loop (AI-EDL): A Conceptual Framework to Bridge AI Capabilities with Classical Educational Theories
by: Yu, Ning, et al.
Published: (2025)

Investigating Language Model Capabilities to Represent and Process Formal Knowledge: A Preliminary Study to Assist Ontology Engineering
by: Akl, Hanna Abi
Published: (2025)