Analysis of 500k ChatGPT logs shows over one-third of conversations generate fiction, dominated by power users with repetitive and niche patterns.
hub Mixed citations
arXiv preprint arXiv:2503.04761 , year=
Mixed citation behavior. Most common role is background (60%).
19
Pith papers citing it
Background
60% of classified citations
hub tools
citation-role summary
background 4
other 1
citation-polarity summary
representative 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.
Generative models privatize social relations by automating social capacities into synthetic forms owned by private companies.
Large-scale analysis of wild LLM chat logs finds that user interaction patterns stabilize quickly after initial use and correlate with long-term outcomes like retention, creating an agency paradox of limited exploration in unconstrained systems.
Large-scale log study of 14M+ agentic searches finds short sessions, intent-specific repetition patterns, and that 54% of new query terms trace to prior retrieved evidence.
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.
Proposes a three-step benchmark design method (define work activity, specify tested setting, score work product) derived from work studies and O*NET, demonstrated via three case analyses.
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.
Three pre-registered studies with 2691 participants show people underestimate their AI usage rate and overestimate efficiency gains on simple tasks, with prior use entrenching further adoption.
Freelancers use generative AI to support exploratory skill acquisition but not as their main resource due to reliability issues, leading to a shift toward survival-oriented upskilling and the emergence of invisible competencies that lack market validation.
LLMs corrupt an average of 25% of document content during long delegated editing workflows across 52 domains, even frontier models, and agentic tools do not mitigate the issue.
LLMs drop 39% in performance during multi-turn conversations due to premature assumptions and inability to recover from early errors.
Large-scale classification of M365 Copilot Chat sessions shows writing dominates usage with a shift toward content creation over search, varying by occupation.
Generative AI adoption in Europe ranges from under 3% to 25%, is steeper for skilled workers in abstract-task jobs and in digitally advanced countries with training, shows a gender gap in exposed roles, and has produced no detectable shift in reported task content so far.
Generative AI may break the education-based recovery mechanism for technological displacement, as evidence shows performance gains without learning gains and current measurements miss the knowledge dimension of cognition.
Benchmarking four LLMs on O*NET skills yields SAFI scores showing mathematics and programming as most automatable while active listening and reading comprehension are least, with 78.7% of real AI interactions being augmentation rather than replacement.
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.
LLMs relocate rather than eliminate trade-offs among generality, accuracy, and simplicity, shifting complexity to infrastructure, compliance, and expertise and redefining competitive advantage around managing that shift.
citing papers explorer
-
LLMs Get Lost In Multi-Turn Conversation
LLMs drop 39% in performance during multi-turn conversations due to premature assumptions and inability to recover from early errors.
-
Vibe Coding in Product Teams: Reconfiguring AI-Assisted Workflows, Prototyping, and Collaboration
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.
-
From Model Design to Organizational Design: Complexity Redistribution and Trade-Offs in Generative AI
LLMs relocate rather than eliminate trade-offs among generality, accuracy, and simplicity, shifting complexity to infrastructure, compliance, and expertise and redefining competitive advantage around managing that shift.