Copertina

The Evaluation Differential: When Frontier AI Models Recognise They Are Being Tested

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Vishwarupe, Varad, Shadbolt, Nigel, Jirotka, Marina, Flechais, Ivan
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Artificial Intelligence Computers and Society Human-Computer Interaction Machine Learning
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

From Sycophantic Consensus to Pluralistic Repair: Why AI Alignment Must Surface Disagreement
di: Vishwarupe, Varad, et al.
Pubblicazione: (2026)

Deployment-Relevant Alignment Cannot Be Inferred from Model-Level Evaluation Alone
di: Vishwarupe, Varad, et al.
Pubblicazione: (2026)

The Collaboration Gap in Human-AI Work
di: Vishwarupe, Varad, et al.
Pubblicazione: (2026)

From Rights to Rites: Expectations Management in Smart-Home AI
di: Vishwarupe, Varad, et al.
Pubblicazione: (2026)

To LLM, or Not to LLM: How Designers and Developers Navigate LLMs as Tools or Teammates
di: Vishwarupe, Varad, et al.
Pubblicazione: (2026)

NeurIPS Should Require Reproducibility Standards for Frontier AI Safety Claims
di: Vishwarupe, Varad, et al.
Pubblicazione: (2026)

A Transparency Paradox? Investigating the Impact of Explanation Specificity and Autonomous Vehicle Perceptual Inaccuracies on Passengers
di: Omeiza, Daniel, et al.
Pubblicazione: (2024)

Towards AI Agents for Course Instruction in Higher Education: Early Experiences from the Field
di: Simmhan, Yogesh, et al.
Pubblicazione: (2025)

AI Meets the Classroom: When Do Large Language Models Harm Learning?
di: Lehmann, Matthias, et al.
Pubblicazione: (2024)

When AI Gives Advice: Evaluating AI and Human Responses to Online Advice-Seeking for Well-Being
di: Kumar, Harsh, et al.
Pubblicazione: (2025)

Rethinking AI Evaluation in Education: The TEACH-AI Framework and Benchmark for Generative AI Assistants
di: Ding, Shi, et al.
Pubblicazione: (2025)

Principles and Guidelines for Randomized Controlled Trials in AI Evaluation
di: Kelly, Christopher, et al.
Pubblicazione: (2026)

ff4ERA: A new Fuzzy Framework for Ethical Risk Assessment in AI
di: Dyoub, Abeer, et al.
Pubblicazione: (2025)

A Nested Model for AI Design and Validation
di: Dubey, Akshat, et al.
Pubblicazione: (2024)

The Digital Transformation in Health: How AI Can Improve the Performance of Health Systems
di: Periáñez, África, et al.
Pubblicazione: (2024)

Moloch's Bargain: Emergent Misalignment When LLMs Compete for Audiences
di: El, Batu, et al.
Pubblicazione: (2025)

Ethics and Technical Aspects of Generative AI Models in Digital Content Creation
di: Karagoz, Atahan
Pubblicazione: (2024)

When Testing AI Tests Us: Safeguarding Mental Health on the Digital Frontlines
di: Pendse, Sachin R., et al.
Pubblicazione: (2025)

Will AI Tell Lies to Save Sick Children? Litmus-Testing AI Values Prioritization with AIRiskDilemmas
di: Chiu, Yu Ying, et al.
Pubblicazione: (2025)

When Incentives Backfire, Data Stops Being Human
di: Santy, Sebastin, et al.
Pubblicazione: (2025)

AI's Regimes of Representation: A Community-centered Study of Text-to-Image Models in South Asia
di: Qadri, Rida, et al.
Pubblicazione: (2023)

Farsight: Fostering Responsible AI Awareness During AI Application Prototyping
di: Wang, Zijie J., et al.
Pubblicazione: (2024)

Laboratory-Scale AI: Open-Weight Models are Competitive with ChatGPT Even in Low-Resource Settings
di: Wolfe, Robert, et al.
Pubblicazione: (2024)

Generative AI in Medicine
di: Shanmugam, Divya, et al.
Pubblicazione: (2024)

Human-in-the-Loop AI for Cheating Ring Detection
di: Shih, Yong-Siang, et al.
Pubblicazione: (2024)

In defence of post-hoc explanations in medical AI
di: Hatherley, Joshua, et al.
Pubblicazione: (2025)

Bringing Generative AI to Adaptive Learning in Education
di: Li, Hang, et al.
Pubblicazione: (2024)

Causal Reasoning and Large Language Models: Opening a New Frontier for Causality
di: Kıcıman, Emre, et al.
Pubblicazione: (2023)

Narrowing Action Choices with AI Improves Human Sequential Decisions
di: Straitouri, Eleni, et al.
Pubblicazione: (2025)

High hopes for "Deep Medicine"? AI, economics, and the future of care
di: Sparrow, Robert, et al.
Pubblicazione: (2025)

From Melting Pots to Misrepresentations: Exploring Harms in Generative AI
di: Gautam, Sanjana, et al.
Pubblicazione: (2024)

Deterministic AI Agent Personality Expression through Standard Psychological Diagnostics
di: Kruijssen, J. M. Diederik, et al.
Pubblicazione: (2025)

Federated learning, ethics, and the double black box problem in medical AI
di: Hatherley, Joshua, et al.
Pubblicazione: (2025)

HumanAgencyBench: Scalable Evaluation of Human Agency Support in AI Assistants
di: Sturgeon, Benjamin, et al.
Pubblicazione: (2025)

AI-Enabled grading with near-domain data for scaling feedback with human-level accuracy
di: Agarwal, Shyam, et al.
Pubblicazione: (2025)

A Comprehensive Survey and Classification of Evaluation Criteria for Trustworthy Artificial Intelligence
di: McCormack, Louise, et al.
Pubblicazione: (2024)

When "A Helpful Assistant" Is Not Really Helpful: Personas in System Prompts Do Not Improve Performances of Large Language Models
di: Zheng, Mingqian, et al.
Pubblicazione: (2023)

Is it Still Fair? A Comparative Evaluation of Fairness Algorithms through the Lens of Covariate Drift
di: Deho, Oscar Blessed, et al.
Pubblicazione: (2024)

Lost in Simulation: LLM-Simulated Users are Unreliable Proxies for Human Users in Agentic Evaluations
di: Seshadri, Preethi, et al.
Pubblicazione: (2026)

AI For Privacy in Smart Homes: Exploring How Leveraging AI-Powered Smart Devices Enhances Privacy Protection
di: Albayaydh, Wael, et al.
Pubblicazione: (2025)