ICML 2025PastFairness & ethics
ICML Workshop on Technical AI Governance (TAIG)
ICML 2025 Workshop TAIG
- Submission deadline
- May 13, 2025, 11:59 UTCimported from OpenReview — check the website for extensions
- Submission portal
- OpenReview
- Notes
- Topics were auto-suggested and may be imprecise — edits welcome.
Accepted papers (45)
Fetched from OpenReview (v2) on 2026-06-10.
A Blueprint for a Secure EU AI Audit Ecosystem
A Conceptual Framework for AI Capability Evaluations
A Taxonomy for Design and Evaluation of Prompt-Based Natural Language Explanations
Acceleration potential in the GPU design-to-manufacturing pipeline
Access Controls Will Solve the Dual-Use Dilemma
AI Benchmarks: Interdisciplinary Issues and Policy Considerations
Attestable Audits: Verifiable AI Safety Benchmarks Using Trusted Execution Environments
CALMA: Context‑Aligned Axes for Language Model Alignment
Compute Requirements for Algorithmic Innovation in Frontier AI Models
Deprecating Benchmarks: Criteria and Framework
Detecting Compute Structuring in AI Governance is likely feasible
Distributed and Decentralised Training: Technical Governance Challenges in a Shifting AI Landscape
Evaluating LLM Agent Adherence to Hierarchical Principles: A Lightweight Benchmark for Verifying AI Safety Plan Components
Expert Survey: Technical AI Safety & Security Research Priorities
Exploring an Agenda on Memorization-based Copyright Verification
Exploring Functional Similarities of Backdoored Models
ExpProof : Operationalizing Explanations for Confidential Models with ZKPs
Fallacies of Data Transparency: Rethinking Nutrition Facts for AI
Fragile by Design: Formalizing Watermarking Tradeoffs via Paraphrasing
From Individual Experience to Collective Evidence: A Reporting-Based Framework for Identifying Systemic Harms
Guaranteeable Memory: An HBM-Based Chiplet for Verifiable AI Workloads
Hardware-Enabled Mechanisms for Verifying Responsible AI Development
In-House Evaluation Is Not Enough: Towards Robust Third-Party Flaw Disclosure for General-Purpose AI
LibVulnWatch: A Deep Assessment Agent System and Leaderboard for Uncovering Hidden Vulnerabilities in Open-Source AI Libraries
LLMs Can Covertly Sandbag On Capability Evaluations Against Chain-of-Thought Monitoring
Locking Open Weight Models with Spectral Deformation
Marginal Risk Relative to What? Distinguishing Baselines in AI Risk Management
Measuring What Matters: A Framework for Evaluating Safety Risks in Real-World LLM Applications
Meek Models Shall Inherit The Earth
Methodological Challenges in Agentic Evaluations of AI Systems
Position: Formal Methods are the Principled Foundation of Safe AI
Position: Generative AI Regulation Can Learn from Social Media Regulation
Practical Principles for AI Cost and Compute Accounting
Probing Evaluation Awareness of Language Models
Proofs of Autonomy: Scalable and Practical Verification of AI Autonomy
Relative Bias: A Comparative Approach for Quantifying Bias in LLMs
Reproducibility: The New Frontier in AI Governance
Robust ML Auditing using Prior Knowledge
Scaling Limits to AI Chip Production
Societal Capacity Assessment Framework: Measuring Advanced AI Implications for Vulnerability, Resilience, and Transformation
Technical Requirements for Halting Dangerous AI Activities
The Strong, weak and benign Goodhart's law. An independence-free and paradigm-agnostic formalisation
Trends in AI Supercomputers
Trends in Frontier AI Model Count: A Forecast to 2028
Watermarking Without Standards Is Not AI Governance