Analysis of 500k ChatGPT logs shows over one-third of conversations generate fiction, dominated by power users with repetitive and niche patterns.
hub
arXiv preprint arXiv:2507.07935 , year=
15 Pith papers cite this work. Polarity classification is still indexing.
15
Pith papers citing it
hub tools
citation-role summary
background 2
citation-polarity summary
roles
background 2representative citing papers
Firms adjust to generative AI by reallocating hiring (52% of exposure decline) and redesigning tasks within jobs (39.5%), with senior roles shifting earlier via reallocation and junior roles using mixed channels.
Platform user selection creates non-classical measurement error in occupational AI exposure measures from logs, with BLS reweighting attenuating employment coefficient estimates by 42-93%.
A new RL Feasibility Index based on task learnability via reinforcement learning diverges from prior AI exposure measures, rating operational jobs like power plant operators as highly feasible while rating creative and interpersonal roles as less so.
Longitudinal analysis of Bing Copilot users shows sticky individual LLM habits, activity-level differences in task complexity and success, and that WildChat is skewed toward power users.
Preregistered behavioral study identifies a speedup illusion where users overestimate time savings from AI assistance on cognitive tasks despite no actual difference in completion times.
The authors propose a retrieval-augmented framework that grounds AI exposure labels for 18,796 O*NET occupation-task pairs in retrieved news and academic abstracts, outperforming zero-shot prompting in 72% of disagreements and aligning better with observed real-world usage.
Large-scale classification of M365 Copilot Chat sessions shows writing dominates usage with a shift toward content creation over search, varying by occupation.
Hiding generative AI use to signal expertise reduces knowledge sharing and transparency among workplace colleagues.
Interviews reveal a four-stage vibe coding workflow that accelerates prototyping while introducing tensions between quick efficiency and reflective design intention, plus asymmetries in trust and ownership.
Machine interpreting should shift from fidelity metrics to three design priorities—agency, grounding, and experience—drawn from interpreting studies to close the usability gap with human-mediated communication.
Language access managers express conditional optimism about AI implementations but emphasize strong risk awareness and the necessity of human oversight and value in translation services.
Survey of Italian adults finds generative AI adoption and capital-enhancing uses stratified by education, age, tech familiarity, and gender, with AI training as key predictor of purposeful engagement.
Chatbot AI systems often fail complex needs while projecting authority, contributing to deskilling, labor displacement, economic concentration, and high environmental costs, so alternative pluralistic and task-specific designs are needed.
citing papers explorer
-
AI Fiction in the Wild
Analysis of 500k ChatGPT logs shows over one-third of conversations generate fiction, dominated by power users with repetitive and niche patterns.
-
Generative AI and the Reorganization of Labor Demand
Firms adjust to generative AI by reallocating hiring (52% of exposure decline) and redesigning tasks within jobs (39.5%), with senior roles shifting earlier via reallocation and junior roles using mixed channels.
-
Who Uses AI? Platform Selection and the Measurement of Occupational AI Exposure
Platform user selection creates non-classical measurement error in occupational AI exposure measures from logs, with BLS reweighting attenuating employment coefficient estimates by 42-93%.
-
What Jobs Can AI Learn? Measuring Exposure by Reinforcement Learning
A new RL Feasibility Index based on task learnability via reinforcement learning diverges from prior AI exposure measures, rating operational jobs like power plant operators as highly feasible while rating creative and interpersonal roles as less so.
-
Adopt $\neq$ Adapt: Longitudinal Analyses of LLM Conversations in the Wild
Longitudinal analysis of Bing Copilot users shows sticky individual LLM habits, activity-level differences in task complexity and success, and that WildChat is skewed toward power users.
-
Cognitive offloading and the speedup illusion in human-AI interaction
Preregistered behavioral study identifies a speedup illusion where users overestimate time savings from AI assistance on cognitive tasks despite no actual difference in completion times.
-
Jobs' AI Exposure Should Be Measured from Evidence, Not Model Priors
The authors propose a retrieval-augmented framework that grounds AI exposure labels for 18,796 O*NET occupation-task pairs in retrieved news and academic abstracts, outperforming zero-shot prompting in 72% of disagreements and aligning better with observed real-world usage.
-
AI in the Enterprise: How People Use M365 Copilot Chat
Large-scale classification of M365 Copilot Chat sessions shows writing dominates usage with a shift toward content creation over search, varying by occupation.
-
"If You're Very Clever, No One Knows You've Used It": The Social Dynamics of Developing Generative AI Literacy in the Workplace
Hiding generative AI use to signal expertise reduces knowledge sharing and transparency among workplace colleagues.
-
Bridging the Usability Gap: Lessons from Interpreting Studies for Machine Interpreting Design
Machine interpreting should shift from fidelity metrics to three design priorities—agency, grounding, and experience—drawn from interpreting studies to close the usability gap with human-mediated communication.
-
AI Technologies in Language Access: Attitudes Towards AI and the Human Value of Language Access Managers
Language access managers express conditional optimism about AI implementations but emphasize strong risk awareness and the necessity of human oversight and value in translation services.
-
What if AI systems weren't chatbots?
Chatbot AI systems often fail complex needs while projecting authority, contributing to deskilling, labor displacement, economic concentration, and high environmental costs, so alternative pluralistic and task-specific designs are needed.
- Measuring and Mitigating Persona Distortions from AI Writing Assistance