Violet Teaming in Practice
A CSET pilot workshop tests violet teaming, a method integrating technical, security and policy views, to govern biodefense-relevant AI capabilities, surfacing non-obvious misuse pathways and implementation constraints.
Research
CSET
Research on AI governance and emerging technology. cset.georgetown.edu
Focus All
Topic All
Source CSET · 26
Date Any time
1 item
1 item
AI Governance at the Frontier
CSET report presents an analytic approach for U.S. policymakers to deconstruct assumptions behind five frontier AI governance proposals and preserve decision-making flexibility under uncertainty.
1 item
The Mechanisms of AI Harm: Lessons Learned from AI Incidents
Analyzes AI Incident Database reports to identify six mechanisms through which AI deployment causes harm, arguing mitigation needs diverse sociotechnical strategies beyond model-based approaches.
1 item
Harmonizing AI Guidance: Distilling Voluntary Standards and Best Practices into a Unified Framework
CSET distills over 7,000 recommended practices from 52 AI standards and guidance reports into a single 258-practice framework for organizations adopting AI, integrating safety and security.
1 item
AI for Military Decision-Making
CSET report giving military leaders a framework for deploying AI decision-support systems, covering scope, context shifts, data biases, automation pitfalls and practical risk mitigations.
1 item
How to Assess the Likelihood of Malicious Use of Advanced AI Systems
A CSET report proposes a framework for policymakers to assess the likelihood of malicious use of advanced AI, covering plausibility, performance, and observed-use evidence.
1 item
AI Incidents: Key Components for a Mandatory Reporting Regime
Proposes standardized key components for a mandatory AI incident reporting regime, covering incident type, harm severity, technical data, affected parties and context, drawing on transportation, healthcare and cybersecurity reporting.
1 item
Governing AI with Existing Authorities
Argues U.S. agencies should first use existing sectoral authorities to govern AI safety, using commercial aviation as a case study to identify gaps needing new legislation.
1 item
Enabling Principles for AI Governance
CSET report arguing that effective AI governance requires tracking AI risks and incidents, building regulator AI literacy, and keeping policies adaptable as AI evolves.
1 item
Putting Teeth into AI Risk Management
A CSET report arguing that federal procurement rules, as used in cybersecurity, can enforce AI risk management and development best practices among government suppliers, and offering recommendations for doing so.
1 item
An Argument for Hybrid AI Incident Reporting
Proposes a federated hybrid framework of mandatory, voluntary and citizen reporting to standardize and aggregate AI incident data, drawing lessons from other sectors' incident databases.
1 item
Comment on NIST RFI Related to the Executive Order Concerning Artificial Intelligence (88 FR 88368)
CSET's comment to NIST compiles recommendations on AI red-teaming, incident criteria, risk management and standards to help implement the AI Executive Order's safety requirements.
1 item
Comment on OMB RFC on Advancing Governance, Innovation, and Risk Management for Agency Use of Artificial Intelligence Draft Memorandum
CSET's comment to OMB on draft federal guidance covering Chief AI Officers, AI risk management and agency AI adoption, arguing for monitoring foreign government AI use and calibrating security practices to offensive uses.
1 item
Repurposing the Wheel: Lessons for AI Standards
A CSET and CNAS workshop report drawing lessons from finance, worker safety, cybersecurity and medical-device standards to inform AI standards development and governance.
1 item
AI Incident Collection: An Observational Study of the Great AI Experiment
CSET explainer defining criteria for AI incident collection and comparing mandatory, voluntary and citizen reporting models for tracking real-world AI harms.
1 item
Adding Structure to AI Harm
CSET introduces a standardized conceptual framework for defining, classifying and tracking real-world harms caused by deployed AI systems, to support risk mitigation analyses.
1 item
Reducing the Risks of Artificial Intelligence for Military Decision Advantage
A CSET policy brief examining how technical AI failures in military decision-making could interact with strategic pressures and human factors to trigger escalation, and how to reduce those risks.
1 item
One Size Does Not Fit All
CSET's research agenda argues AI assessment must cover diverse products and tools, and proposes studying testing, standards, auditing and certification to reduce AI safety and trust risks.
1 item
Comment to NIST on the AI Risk Management Framework
CSET's comment on NIST's second draft of the AI Risk Management Framework praises revisions to audience, risk definition and the Govern function, and flags interactions among AI systems as a future risk.
1 item
Exploring Clusters of Research in Three Areas of AI Safety
A CSET bibliometric brief mapping citation clusters of research on AI safety robustness, interpretability and reward learning, describing growth trends, active countries and influential papers.
1 item
Key Concepts in AI Safety: Specification in Machine Learning
A CSET explainer introducing specification as a core AI safety problem: conveying designers' intentions to machine learning systems, including reward misspecification and principal-agent analogies.
1 item
Classifying AI Systems
CSET brief developing and testing frameworks to classify AI systems along policy-relevant dimensions, comparing over 1,800 classifications to support risk assessment, governance and system management.
1 item
AI Accidents: An Emerging Threat
CSET policy brief classifying AI accident types (robustness, specification, assurance failures) with hypothetical scenarios and recommending policy actions to reduce accident risks from deployed ML systems.
2 items
Key Concepts in AI Safety: Interpretability in Machine Learning
CSET primer framing interpretability as a means to achieve assurance in ML systems within the AI safety agenda, explaining how understanding and designing interpretable models helps catch unintended behaviour.
Key Concepts in AI Safety: An Overview
A CSET primer that introduces robustness, assurance, and specification as three categories of AI safety problems in machine learning, with links to follow-up briefs.
1 item
AI Verification
Proposes concepts and technical mechanisms for verifying states' compliance with international arms control rules on military AI, including inspections and continuous verification.