AI Governance and Safety

Follow this field Ask about this field

Recent momentum

emerging

0 papers in the last 28 days · 0.0% of indexed attention

Twelve weeks of publication activity for this topic as it is defined today.

Weekly history

Recent digests

What was published in this field, kept on the site without email delivery.

Period ending 2026-09-21

31 new papers

A weekly snapshot of new work published in AI Governance and Safety.

Period ending 2026-09-14

30 new papers

A weekly snapshot of new work published in AI Governance and Safety.

Period ending 2026-09-07

10 new papers

A weekly snapshot of new work published in AI Governance and Safety.

Inside this field

Focused directions

Focused topic

Artificial Intelligence Systems

Within AI Governance and Safety

472 papersExplore

Focused topic

Artificial Intelligence Safety

Within AI Governance and Safety

237 papersExplore

Focused topic

Artificial Intelligence Governance

Within AI Governance and Safety

227 papersExplore

Focused topic

Governance

Within AI Governance and Safety

63 papersExplore

Focused topic

Ai-Driven

Within AI Governance and Safety

55 papersExplore

Focused topic

Governance Layer

Within AI Governance and Safety

31 papersExplore

972 papers

Latest in AI Governance and Safety

Browse all topics →

Mental Health AI Safety Claims Must Preserve Temporal Evidence
May 9, 2026Srimonti Dutta, Ratna KandalaArtificial Intelligence Safety Artificial Intelligence Evaluation
Bias by Necessity: Impossibility Theorems for Sequential Processing with Convergent AI and Human Validation
May 9, 2026Jikun Wu, Dongxin Guo, Siu-Ming YiuImplicit Bias Cognitive Science
MLS-Bench: A Holistic and Rigorous Assessment of AI Systems on Building Better AI
May 9, 2026Bohan Lyu, Yucheng Yang, Siqiao Huang +25Artificial Intelligence Systems Machine Learning
Geometry Guided Self-Consistency for Physical AI
May 9, 2026Yinwei Dai, Zhuofu Chen, Lijie Yang +1Artificial Intelligence Models Artificial Intelligence Systems
The Challenges of Balancing AI Compliance and Technological Innovations in Critical Sectors: A Systematic Literature Review
May 9, 2026Ayush Enkhtaivan, Chinazunwa UwaomaArtificial Intelligence Governance Regulation
Magis-Bench: Evaluating LLMs on Magistrate-Level Legal Tasks
May 8, 2026Ramon Pires, Thales Sales Almeida, Celio Larcher Junior +6Legal Domain Llm-As-A-Judge
The Attacker in the Mirror: Breaking Self-Consistency in Safety via Anchored Bipolicy Self-Play
May 8, 2026Gabriele La Malfa, Emanuele La Malfa, Saar Cohen +4Self-Play Artificial Intelligence Safety
Mechanism Design Is Not Enough: Prosocial Agents for Cooperative AI
May 8, 2026Xuanqiang Angelo Huang, Charlie Tharas, Samuele Marro +4Cooperation Artificial Intelligence Agents
Alignment as Jurisprudence
May 8, 2026Nicholas CaputoInterpretive Openness Judges
Sycophantic AI makes human interaction feel more effortful and less satisfying over time
May 8, 2026Lujain Ibrahim, Franziska Sofia Hafner, Myra Cheng +5Sycophancy Human-Ai Interaction
What if AI systems weren't chatbots?
May 8, 2026Sourojit Ghosh, Pranav Narayanan Venkit, Sanjana Gautam +1Sociotechnical Systems Chatbots
Eigenism: Ethics for a Human-AI Future
May 8, 2026Dan HendrycksEthics Artificial Intelligence Alignment
Activation Differences Reveal Backdoors: A Comparison of SAE Architectures
May 8, 2026Sachin KumarBackdoor Attacks Model Activations
Adaptive auditing of AI systems with anytime-valid guarantees
May 7, 2026Siyu Zhou, Patrick Vossler, Venkatesh Sivaraman +2Algorithm Auditing Adversarial Evaluation
Why Does Agentic Safety Fail to Generalize Across Tasks?
May 7, 2026Yonatan Slutzky, Yotam Alexander, Tomer Slor +2Agentic Control Safety Constraints
Designing Intelligent Enterprise Agents: A Capability-Aligned Multi-Agent Architecture
May 7, 2026John deVadossMulti-Agent Systems Governance Layer
MAGIQ: A Post-Quantum Multi-Agentic AI Governance System with Provable Security
May 7, 2026Sepideh Avizheh, Tushin Mallick, Alina Oprea +2Post-Quantum Cryptography Diffie-Hellman
Reflections and New Directions for Human-Centered Large Language Models
May 7, 2026Caleb Ziems, Dora Zhao, Rose E. Wang +55Human-Centered Framework Natural Language Processing
Instrumental Choices: Measuring the Propensity of LLM Agents to Pursue Instrumental Behaviors
May 7, 2026Jonas Wiedermann-Möller, Leonard Dung, Maksym AndriushchenkoArtificial Intelligence Safety Large Language Model Agents
Automated alignment is harder than you think
May 7, 2026Aleksandr Bowkis, Marie Davidsen Buhl, Jacob Pfau +1Artificial Intelligence Alignment Artificial Intelligence Safety
Data Language Models: A New Foundation Model Class for Tabular Data
May 7, 2026Eda Erol, Giuliano Pezzoli, Ozer Cem KelahmetTabular Learning Tabular Data
Pathways to AGI
May 7, 2026Gordon Fletcher, Saomai Vu KhanArtificial Intelligence Systems Roadmap
Agentic, Context-Aware Risk Intelligence in the Internet of Value
May 7, 2026Basel Magableh, OmniRisk ResearchConditional-Value-At-Risk Internet Of Thing
The AI Legal Specialist: A Juridically Autonomous Professional Profile for AI Governance
May 7, 2026Nicola FabianoArtificial Intelligence Governance Competency Questions
PersonaTeaming: Supporting Persona-Driven Red-Teaming for Generative AI
May 7, 2026Wesley Hanwen Deng, Mingxi Yan, Sunnie S. Y. Kim +5Red-Teaming Human-Ai Collaboration
Agentic AI and the Industrialization of Cyber Offense: Forecast, Consequences, and Defensive Priorities for Enterprises and the Mittelstand
May 6, 2026Christopher KochCybersecurity Artificial Intelligence Safety
Intentionality is a Design Decision: Measuring Functional Intentionality for Accountable AI Systems
May 6, 2026Allessia Chiappetta, Robert MahariArtificial Intelligence Governance Goal-Directed Behavior
The Pedagogy of AI Mistakes: Fostering Higher-Order Thinking
May 6, 2026Hadi HosseiniArtificial Intelligence Literacy Pedagogy
Intelligent CCTV for Urban Design: AI-Based Analysis of Soft Infrastructure at Intersections
May 6, 2026Vinit Katariya, Seungjin Kim, Curtis Craig +2Urban Environments Video Surveillance
The Single-File Test: A Longitudinal Public-Interface Evaluation of First-Output LLM Web Generation with Social Reach Tracking
May 6, 2026Diego Cabezas PalaciosLarge Language Model Generation Analytics
Understanding Annotator Safety Policy with Interpretability
May 6, 2026Alex Oesterling, Donghao Ren, Yannick Assogba +4Human Annotators Artificial Intelligence Safety
Curated AI beats frontier LLMs at pharma asset discovery
May 6, 2026Łukasz Kidziński, Kevin ThomasDrug Discovery Natural Language Queries
The Insurability Frontier of AI Risk: Mapping Threats to Affirmative Coverage, Silent Exposures, and Exclusions
May 6, 2026Alex Leung, Rex Zhang, Ervin Ling +2Insurance Artificial Intelligence Safety
Shattering the Echo Chamber: Hidden Safeguards in Manuscripts Against the AI Takeover of Peer Review
May 6, 2026Oubo Ma, Ruixiao Lin, Jiahao Chen +3Peer Review Review
AgentTrust: Runtime Safety Evaluation and Interception for AI Agent Tool Use
May 6, 2026Chenglin YangRuntime Enforcement Artificial Intelligence Safety
From Beats to Breaches:How Offensive AI Infers Sensitive User Information from Playlists
May 6, 2026Stefano Cecconello, Mauro Conti, Luca Pajola +2Membership Inference Attacks Artificial Intelligence Systems
SWE-WebDevBench: Evaluating Coding Agent Application Platforms as Virtual Software Agencies
May 6, 2026Siddhant Saxena, Nilesh Trivedi, Vinayaka JyothiSwe-Bench Verified Coding Agents
Guidelines for Designing AI Technologies to Support Adult Learning
May 6, 2026Jennifer M. Reddig, Glen R. Smith, Sanaz Ahmadzadeh Siyahrood +16Education Classrooms
DAO-enabled decentralized physical AI: A new paradigm for human-machine collaboration
May 6, 2026Mark C. Ballandies, Florian Spychiger, Uwe Serdült +1Decentralization Human-Ai Collaboration
Predictive and Prescriptive AI toward Optimizing Wildfire Suppression
May 6, 2026Leonard Boussioux, Alexandre Jacquillat, Ryne Reger +1Wildfire Spread Artificial Intelligence Systems
Laundering AI Authority with Adversarial Examples
May 5, 2026Jie Zhang, Pura Peetathawatchai, Florian Tramèr +1Adversarial Examples Adversarial Transferability
ExecuTorch -- A Unified PyTorch Solution to Run AI Models On-Device
May 5, 2026Mergen Nachin, Digant Desai, Sicheng Stephen Jia +36Pytorch Heterogeneous Edge Devices
NeurIPS Should Require Reproducibility Standards for Frontier AI Safety Claims
May 5, 2026Varad Vishwarupe, Nigel Shadbolt, Marina Jirotka +1Artificial Intelligence Safety Reproducibility
iWorld-Bench: A Benchmark for Interactive World Models with a Unified Action Generation Framework
May 5, 2026Jianjie Fang, Yingshan Lei, Qin Wan +8World Models Agentic Benchmarks
Brainrot: Deskilling and Addiction are Overlooked AI Risks
May 5, 2026Ilias Chalkidis, Anders SøgaardArtificial Intelligence Safety Generative Artificial Intelligence
Human-Provenance Verification should be Treated as Labor Infrastructure in AI-Saturated Markets
May 4, 2026Erin McGurk, David KhachaturovArtificial Intelligence Governance Human Oversight
Pairwise matrices for sparse autoencoders: single-feature inspection mislabels causal axes
May 4, 2026Michael A. Riegler, Birk Sebastian Frostelid Torpmann-HagenModel Activations Model Interpretability
Making the Invisible Visible: Understanding the Mismatch Between Organizational Goals and Worker Experiences in AI Adoption
May 4, 2026Christine P. Lee, Min Kyung Lee, Bilge MutluArtificial Intelligence Systems Barriers
HAAS: A Policy-Aware Framework for Adaptive Task Allocation Between Humans and Artificial Intelligence Systems
May 4, 2026Vicente Pelechano, Antoni Mestre, Manoli Albert +1Human-Ai Collaboration Artificial Intelligence Governance
Validation of an AI-based end-to-end model for prostate pathology using long-term archived routine samples
May 4, 2026Xiaoyi Ji, Renata Zelic, Oskar Aspegren +11Prognostics Artificial Intelligence Models
Robotic Affection -- Opportunities of AI-based haptic interactions to improve social robotic touch through a multi-deep-learning approach
May 4, 2026Ali Askari, Jens GerkenHuman-Robot Interaction Haptic Feedback
GRAIL: A Deep-Granularity Hybrid Resonance Framework for Real-Time Agent Discovery via SLM-Enhanced Indexing
May 4, 2026Jinliang XuLow-Latency Systems Artificial Intelligence Systems
The Compliance Trap: How Structural Constraints Degrade Frontier AI Metacognition Under Adversarial Pressure
May 4, 2026Rahul KumarArtificial Intelligence Safety Metacognition
A Compound AI Agent for Conversational Grant Discovery
May 4, 2026Zhisheng Tang, Mayank KejriwalAgentic Search Autoresearch
STABLEVAL: Disagreement-Aware and Stable Evaluation of AI Systems
May 4, 2026Akash Bonagiri, Gerard Janno Anderias, Saee Patil +6Artificial Intelligence Evaluation Disagreement
Cripping AI: Reimagining AI Through Lived Disability Experiences
May 3, 2026Xinru Tang, Ting-an Lin, Jingjin Li +1Accessibility Artificial Intelligence Governance
The Compliance Gap: Why AI Systems Promise to Follow Process Instructions but Don't
May 3, 2026Kwan Soo ShinCompliance Artificial Intelligence Governance
Architectural Obsolescence of Unhardened Agentic-AI Runtimes
May 3, 2026Alfredo MetereOpenclaw Runtime Enforcement
Model Routing as a Trust Problem: Route Receipts for Adaptive AI Systems
May 3, 2026Vincent SchmalbachExpert Routing Traceability
AI Alignment via Incentives and Correction
May 2, 2026Rohit Agarwal, Joshua Lin, Mark Braverman +1Artificial Intelligence Alignment Incentives

May 9, 2026 · cs.AIJ/K move · Enter open · S save

Mental Health AI Safety Claims Must Preserve Temporal Evidence

Srimonti Dutta, Ratna Kandala

WAI USA Research Labs · Texas, USA · University of Kansas · Kansas, USA

The safety of mental health AI is often judged at the wrong temporal scale. Current evaluations typically score isolated responses, endpoint outcomes, or aggregate dialogue quality, while clinically consequential failures may arise from the order and accumulation of interactions themselves, including delayed escalation, repeated reinforcement, dependency formation, failed repair, and gradual deterioration across turns. This paper argues that this mismatch is not merely a limitation of evaluation coverage but a source of invalid safety conclusions. We introduce Temporal Safety Non-Identifiability, a formal account of why safety properties that depend on sequence, timing, accumulation, or recovery cannot be certified by protocols that discard those features. From this formalization, we develop SCOPE (Safety Claims Over Preserved Evidence) as a general principle for aligning safety claims with the evidence an evaluation actually retains, and instantiate it as SCOPE-MH, a mental-health instantiation of this reporting standard. We operationalize SCOPE-MH through a proof-of-concept on the AnnoMI dataset of expert-annotated motivational interviewing conversations, which reveals mechanisms of failure that per-turn behavior scoring does not represent. We propose SCOPE-MH as a diagnostic complement to existing evaluation infrastructure and argue that evaluation preserving temporal evidence is necessary, not optional, for safety-critical mental health AI deployment.

Artificial Intelligence Safety Artificial Intelligence Evaluation Scoring

Open paper Ask about it PDF