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
8Unique collaborators
2/2Semantic Scholar citation coverage
Publication span: 2026. Corpus fetch span: 2026.
Identity provenance
Provider IDs
- Semantic Scholar:
2357014651
ORCID evidence
No valid ORCID is stored.
Observed aliases (1)
- James Zou (semantic scholar, provider refresh)
Topics and outcomes in this view
Assessment themes
- Human Ai Collab: 2 papers
- Governance: 1 paper
- Productivity: 1 paper
Claim outcomes
- Research Productivity: 1 paper
- Team Performance: 1 paper
- Output Quality: 1 paper
- Decision Quality: 1 paper
- Task Allocation: 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 agents mirror human 'forking paths', reproducing 72% of the ideological gap seen across 42 human teams and often producing opposing conclusions from identical data; an Agentic Bootstrap and the new m-value reveal that many published analyses sit far out on the tails of the plausible analysis distribution, suggesting selective exploration more than outright analytic error.arxiv | James Zou provider id |
2026-07-01 | 2 |
| Self-organizing LLM teams systematically underuse their experts and lag behind the best member, suffering performance drops of up to 41% on ML benchmarks; teams average expert and non-expert views—a consensus-seeking bias that grows with size but makes them more robust to adversaries.arxiv | James Zou provider id |
2026-02-01 | 11 |
Citation observation summary
Semantic Scholar supplied counts for 2 of 2 papers in this view; 0 are missing. The observed paper counts sum to 13 cumulative citations. This is a coverage summary, not an author score or h-index.