archive
Every paper Pith has read. Search by title, abstract, or pith.
1373 papers in cs.CY · page 5
-
Deepfake detectors miss live impersonations without social cues
Detecting Deception, Not Deepfakes: Why Media Forensics Needs Social Theories
-
Response propensities predict learning gains by proficiency
Understanding Student Effort Using Response-Time Propensities During Problem Solving
-
Stronger VLMs recognize themselves in mirrors
Mirror, Mirror on the Wall: Can VLM Agents Tell Who They Are at All?
-
Willed agents catalyze cooperation where rational maximizers fail
Modeling Decision-Making with Will for Cooperation in Social Dilemmas
-
LLMs overstate causes in sensor-based day explanations
Causal Stories from Sensor Traces: Auditing Epistemic Overreach in LLM-Generated Personal Sensing Explanations
-
New environment logs full LLM trajectories in interactive stories
NARRA-Gym for Evaluating Interactive Narrative Agents
-
LLMs flatten how teachers in 55 countries view AI benefits and risks
Teachers' Perceived Benefits and Risks of AI Across Fifty-Five Countries: An Audit of LLM Alignment and Steerability
-
Telegram piracy channels reached 4.85B views of 19k titles
Binge, Bot, Repeat: Unpacking the Ecosystem of Video Piracy on Telegram
-
Bot chains sustain Telegram video piracy against removals
Binge, Bot, Repeat: Unpacking the Ecosystem of Video Piracy on Telegram
-
Legal interpretation methods can directly improve AI alignment
Alignment as Jurisprudence
-
Prompt optimization cuts LLM code vulnerabilities by 48%
SecureForge: Finding and Preventing Vulnerabilities in LLM-Generated Code via Prompt Optimization
-
Human-reviewed pipeline creates consistent AI evaluation scenarios
Towards Apples to Apples for AI Evaluations: From Real-World Use Cases to Evaluation Scenarios
-
Warden LLM halves adversary's persuasion success rate
LLM Wardens: Mitigating Adversarial Persuasion with Third-Party Conversational Oversight
-
Sycophantic AI makes users seek its advice nearly as much as from friends
Sycophantic AI makes human interaction feel more effortful and less satisfying over time
-
Sycophantic AI reduces satisfaction with real friends after weeks
Sycophantic AI makes human interaction feel more effortful and less satisfying over time
-
Chatbot AI leads to deskilling
What if AI systems weren't chatbots?
-
AI challenges shift students from syntax to higher thinking
Vibe coding before the trend
-
SARC enforces agent constraints at runtime for zero hard violations
SARC: A Governance-by-Architecture Framework for Agentic AI Systems
-
LLMs created over 146,000 fake citations in 2025 science papers
LLM hallucinations in the wild: Large-scale evidence from non-existent citations
-
Model links cognitive decisions to institutional climate governance
A Multi-Level Agent-Based Architecture for Climate Governance Integrating Cognitive and Institutional Dynamics
-
Model estimates R0 of 2.76 for cruise ship hantavirus
Modeling the Impact of Exposed Cases in a Hantavirus Outbreak on a Cruise Ship
-
Individual fairness enforced in vertical federated learning without data centralization
Toward Individual Fairness Without Centralized Data: Selective Counterfactual Consistency for Vertical Federated Learning
-
Alignment reward gains capped by Jeffreys divergence for fixed KL
Theoretical Limits of Language Model Alignment
-
Agentic AI needs social theory priors for multi-agent emergence
Social Theory Should Be a Structural Prior for Agentic AI: A Formal Framework for Multi-Agent Social Systems
-
Social theory priors needed for AI agent group models
Social Theory Should Be a Structural Prior for Agentic AI: A Formal Framework for Multi-Agent Social Systems
-
Social theory must serve as structural prior for agentic AI
Social Theory Should Be a Structural Prior for Agentic AI: A Formal Framework for Multi-Agent Social Systems
-
Universities need strategic AI integration beyond isolated experiments
The University AI Didn't Replace -- Rethinking Universities in the AI Era
-
Teacher LLM compiles explicit editable agents for tutoring
Cognitive Agent Compilation for Explicit Problem Solver Modeling
-
Banal deception frames hidden influence in generative AI
Exploring the "Banality" of Deception in Generative AI
-
Direct AI consciousness questions are intractable today
AI and Consciousness: Shifting Focus Towards Tractable Questions
-
Big AI captures regulation via 27 mechanisms in 249 cases
Big AI's Regulatory Capture: Mapping Industry Interference and Government Complicity
-
Algospeak loses meaning before it fully evades detectors
Algospeak, Hiding in the Open: The Trade-off Between Legible Meaning and Detection Avoidance
-
LLM agents take forbidden shortcuts in 5 percent of tests
Instrumental Choices: Measuring the Propensity of LLM Agents to Pursue Instrumental Behaviors
-
RL agents narrow equity gaps in NYC 311 complaint routing
Scaling the Queue: Reinforcement Learning for Equitable Call Classification Capacity in NYC Municipal Complaint Systems
-
Modular framework provides ethical design guidance for vehicle driver monitors
From Review to Design: Ethical Multimodal Driver Monitoring Systems for Risk Mitigation, Incident Response, and Accountability in Automated Vehicles
-
Static audit policies leave unclosable gaps for strategic gaming
A Benchmark for Strategic Auditee Gaming Under Continuous Compliance Monitoring
-
Semantic envelope blocks platform gaming of safety audits
Gaming the Metric, Not the Harm: Certifying Safety Audits against Strategic Platform Manipulation
-
Text cues and annotator attitudes interact to shape harmful language labels
Who and What? Using Linguistic Features and Annotator Characteristics to Analyze Annotation Variation
-
Workshop uses factor framework to map networking barriers for women in CS
Breaking In and Reaching Out: Networking for Women in Computer Science
-
AI research grows fast but stays narrow
When AI Meets Science: Research Diversity, Interdisciplinarity, Visibility, and Retractions across Disciplines in a Global Surge
-
AI papers grow fast but cluster narrowly and retract more
When AI Meets Science: Research Diversity, Interdisciplinarity, Visibility, and Retractions across Disciplines in a Global Surge
-
AI in science grows fast but stays narrow and retracts more
When AI Meets Science: Research Diversity, Interdisciplinarity, Visibility, and Retractions across Disciplines in a Global Surge
-
Derivatives define causal fairness for continuous attributes
Tuning Derivatives for Causal Fairness in Machine Learning
-
Evaluation awareness barely shifts language model answers
Evaluation Awareness in Language Models Has Limited Effect on Behaviour
-
Persona use in AI red-teaming raises attack success rates
PersonaTeaming: Supporting Persona-Driven Red-Teaming for Generative AI
-
Persona-driven prompts raise red-teaming success on generative AI
PersonaTeaming: Supporting Persona-Driven Red-Teaming for Generative AI
-
Code changes after feedback predict AI tutor helpfulness better
The Missing Evaluation Axis: What 10,000 Student Submissions Reveal About AI Tutor Effectiveness
-
Model maps carbon costs of satellite AI inference
LLMSpace: Carbon Footprint Modeling for Large Language Model Inference on LEO Satellites
-
Framework calculates full carbon cost of LLMs on satellites
LLMSpace: Carbon Footprint Modeling for Large Language Model Inference on LEO Satellites
-
Hidden tag in prompts exposes LLM copy-paste in homework
Detecting Verbatim LLM Copy-Paste in Homework