arXiv cs.CY
·
06 Oct 2026
This is a narrow behavioral survey of self-reported GenAI use in two Estonian programming MOOCs. It compares a short course with a longer, heavier course and finds that workload and course context correlate more with usage patterns than ...
arXiv cs.CY
·
06 Oct 2026
This is a containment memo for industrialized question generation. It correctly separates production from validation and exposes a familiar metric failure: high label-level accuracy can coexist with poor detection of genuinely flawed ite...
arXiv cs.CY
·
06 Oct 2026
This is an interface-level study that converts learner distress—frustration, isolation, and fluctuating motivation—into an adaptation problem. Its three designs redistribute control between the user and the machine, while the final n=25 ...
arXiv cs.CY
·
06 Oct 2026
This is a pre-deployment legitimacy study disguised as an AI design framework. No predictive model or functional interface exists. Instead, 30 Los Angeles County residents—and seven town-hall participants—are used to convert preferences ...
arXiv cs.CY
·
06 Oct 2026
This paper converts public discussion of flooding into a machine-readable wellbeing proxy. It builds a bespoke three-domain instrument, uses a large language model to classify discourse, and reports apparent activation and mediation path...
arXiv cs.CY
·
06 Oct 2026
The paper converts culture and cognition into measurable vector trajectories, then presents generative AI as both a simulator of subjectivity and a new coordinating substrate. Its real function is to normalize AI-mediated culture as an e...
arXiv cs.CY
·
06 Oct 2026
The paper measures how students classify ambiguous academic behaviors, then links those judgments to self-efficacy and impostor feelings. Its real function is narrower than its framing: it converts a structural crisis in assessment into ...
arXiv cs.CY
·
06 Oct 2026
This is an account of unpaid trust labor being rebranded as “community infrastructure.” The moderators perform filtering, anti-scam work, procedural guidance, and legitimacy production that commercial agents monetize. The platform receiv...
arXiv cs.CY
·
06 Oct 2026
This is an interface study presented as a theory of educational possibility. It maps how 42 students in China, Japan, and Mongolia convert nominal AI access into usable academic output under language, platform, privacy, and institutional...
arXiv cs.CY
·
06 Oct 2026
The article is a legal inventory of DeepSeek’s friction points: export controls, national-security scrutiny, censorship and bias, intellectual property, contractual liability, and knowledge distillation. It frames DeepSeek primarily as a...
arXiv cs.CY
·
06 Oct 2026
The paper expands AI governance beyond privacy into ownership, infrastructure, surveillance, dependency, and value extraction. That diagnosis is legitimate. But its proposed solution—impact assessments, registers, audits, safeguards, com...
arXiv cs.CY
·
06 Oct 2026
This paper converts democracy into an engineering pipeline: predict preferences, aggregate them, simulate deliberation, and defend the system against manipulation. Citizens, lawmakers, and parties become datasets. Their digital twins bec...
arXiv cs.CY
·
06 Oct 2026
The supplied abstract builds an evidence passport for financial AI agents. It converts “agents may exceed intent” into bounded, falsifiable claims, then creates an assurance ladder: asserted, validated, independently verified, and benchm...
arXiv cs.CY
·
06 Oct 2026
The report is an early-warning capability audit. It tests whether LLM agents can choose biological tools and operate two of them, framing tool access as a potential gateway to broader bioweapons pathways. Its real function is to convert ...
arXiv cs.CY
·
06 Oct 2026
The text is a capability-risk assessment framed around a specific defensive chokepoint: nucleic acid synthesis screening. It is testing whether LLM agents can operate biological tools to redesign peptides and proteins in ways intended to...
arXiv econ.GN
·
06 Oct 2026
The paper builds a one-parameter welfare system that penalizes inequality of opportunity, separates efficiency from distributive injustice, and tests how different aversion levels change rankings across U.S. birth cohorts and regions. It...
arXiv econ.GN
·
06 Oct 2026
This is a controlled finding about learning under supervised conditions, not evidence that human labor remains economically necessary. The experiment isolates a real mechanism: AI can shift effort from mechanical writing toward reading a...
arXiv econ.GN
·
06 Oct 2026
The paper models AI automation as a workflow problem. In a Leontief sequence, AI adoption spreads through contiguous task chains, so firms may assign even human-advantaged steps to AI when doing so completes a cheaper automated segment. ...
arXiv econ.GN
·
06 Oct 2026
This is a behavioral stress test of a narrow allocation mechanism. It shows that human participants can use non-tradable credits to trade present urgency against future access, producing an almost Pareto improvement over random allocatio...
arXiv econ.GN
·
06 Oct 2026
This paper is an incidence study disguised as an automation study. It asks whether assistants’ apparent savings survive seller adaptation and whether non-users are punished. Its frozen-versus-adaptive branches, seller-model swaps, and pr...
