Direction, evidence grade, and study type are AI-generated labels (gpt-5-mini), not human-verified. Syntheses are LLM-written. "Tensions" are machine-detected candidates, not confirmed contradictions. A research-acceleration tool, not peer review.
How this is built →
2Distinct papers
44Unique collaborators
2/2Semantic Scholar citation coverage
Publication span: 2026. Corpus fetch span: 2026.
Identity provenance
Provider IDs
- Semantic Scholar:
2302614848
ORCID evidence
No valid ORCID is stored.
Observed aliases (1)
- Di-Yi Yang (semantic scholar, provider refresh)
Topics and outcomes in this view
Assessment themes
- Human Ai Collab: 2 papers
- Productivity: 2 papers
- Adoption: 1 paper
- Governance: 1 paper
- Labor Markets: 1 paper
- Org Design: 1 paper
- Skills Training: 1 paper
Claim outcomes
- Output Quality: 1 paper
- Task Allocation: 1 paper
- Adoption Rate: 1 paper
- Automation Exposure: 1 paper
- Error Rate: 1 paper
- Firm Productivity: 1 paper
- Governance And Regulation: 1 paper
- Hiring: 1 paper
- Organizational Efficiency: 1 paper
- Other: 1 paper
- Skill Acquisition: 1 paper
- Skill Obsolescence: 1 paper
Papers in the Semantic Scholar view
Latest stored Semantic Scholar author observations only. Citation counts below are from the same provider and are not combined with other services.
Scroll the table horizontally to see every column.
| Paper | Author evidence | Date | Provider citations |
|---|---|---|---|
| AI only augments work when entire workflows—not just isolated tasks—are redesigned to deliver durable net value, preserve meaningful human control and accountability, and sustain learning and career pathways; otherwise short-term productivity gains risk eroding oversight, skills, and job quality.arxiv | Di-Yi Yang provider id |
2026-09-11 | 0 |
| Short sequences of human–agent interaction materially raise agent performance: in experiments with 30 expert users on writing and visual-creation tasks, test-time context and weight adaptation lift agent solo success by about 4.5–20.9% within tens of sessions. An evolving verifier that crystallizes user criteria also flags 16–22% more failures than rubrics from LMs or humans alone, and some personalized behaviors generalize across users.arxiv | Di-Yi Yang provider id |
2026-09-03 | 0 |
Citation observation summary
Semantic Scholar supplied counts for 2 of 2 papers in this view; 0 are missing. The observed paper counts sum to 0 cumulative citations. This is a coverage summary, not an author score or h-index.