In a blind review by 79 NLP researchers, research ideas from a language-model agent were rated more novel than ideas from experts (5.64 against 4.84 on a 10-point scale) and slightly weaker on feasibility; but of 4,000 seed ideas the agent generated per topic, only about 200 were not duplicates.

SocioLens project,

Claim LD09-SL-C19 (v1.0)

Provenance and status

Claim type: Estimate
Publication status: Eligible for publication, with qualifiers: source position
Verification state: Checked against the primary source; the source is a preprint. Who checked this claim is not recorded for it; claim checks so far were AI-led, and no person has checked claims against their sources.
Does not license: That the ideas would work when carried out (see LD09-SL-C20)
Limitations: One field; ideas judged before execution; the agent runs on an Anthropic model and this debate was drafted by an Anthropic model; read in arXiv version 1 (6 September 2024), which predates the ICLR 2025 acceptance
Source (as cited): Si, Yang & Hashimoto, “Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers,” ICLR 2025, arXiv:2409.04109
Source type: Peer-reviewed conference paper (recorded as “Peer-reviewed conference paper”)
Institutional position: Independent academic (Stanford University); the idea-generating agent runs on an Anthropic model
Maturity of the evidence: Emerging evidence
Recorded date: 15 October 2026
Currentness / correction: This is the current version. No correction or concern is recorded on its sources.

Sources this version cites

Si, Yang & Hashimoto, “Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers,” ICLR 2025, arXiv:2409.04109 (Peer-reviewed conference paper)
Read the original

Version

This page shows version v1.0. Its record never changes, even when newer evidence arrives; the site may add a note to the page or withhold a figure not cleared for public use.

This pinned version is also the current version.

Versions: v1.0 · current

Raw record: every recorded field, with its coded values

All 25 canonical Claim fields, verbatim, with the coded values derived from them. The plain-language labels above are read from these.

Claim type (coded): ESTIMATE (as printed: “ESTIMATE”)
Publication status (coded): ELIGIBLE (SOURCE_POSITION) (as printed: “ELIGIBLE WITH SOURCE-POSITION QUALIFIER”)
Verification state (coded): CONTENT_VERIFIED_PRIMARY (PREPRINT)
Source type (coded): PEER_REVIEWED_CONFERENCE_PAPER
Maturity (coded): EMERGING

#	Canonical field	Canonical value (verbatim)	Normalized (Claim Value Vocabulary v1.0.2)
1	CLAIM ID	LD09-SL-C19	—
2	PUBLIC-FACING CLAIM	In a blind review by 79 NLP researchers, research ideas from a language-model agent were rated more novel than ideas from experts (5.64 against 4.84 on a 10-point scale) and slightly weaker on feasibility; but of 4,000 seed ideas the agent generated per topic, only about 200 were not duplicates.	—
3	CLAIM TYPE	ESTIMATE	base_types: ESTIMATE
4	RESEARCH STRAND	B — Creative work, writing and ideas	—
5	MATURITY	EMERGING	states: EMERGING
6	SOURCE	Si, Yang & Hashimoto, “Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers,” ICLR 2025, arXiv:2409.04109	source_ids: SRC-SL-SI-YANG-HASHIMOTO-ICLR-2025
7	SOURCE TYPE	Peer-reviewed conference paper	classes: PEER_REVIEWED_CONFERENCE_PAPER
8	INSTITUTIONAL POSITION	Independent academic (Stanford University); the idea-generating agent runs on an Anthropic model	—
9	SCOPE	Research ideas on prompting in natural language processing	—
10	DESIGN	Blind comparison of expert-written and agent-generated ideas, with a duplication analysis of the agent’s output	—
11	SAMPLE / DATASET	49 ideas per condition; 79 expert reviewers	—
12	MODEL / TOOL / VERSION	An idea-generation agent built on Claude 3.5 Sonnet (June 2024)	—
13	COMPARATOR	Ideas written by NLP experts	—
14	MAIN RESULT	Novelty 5.64 (AI) against 4.84 (human), p < 0.05; about 200 non-duplicate ideas among 4,000 generated per topic (similarity threshold 0.8), with non-duplicates plateauing as generation scales	—
15	INFERENTIAL TARGET	Whether AI-generated research ideas are better individually and how varied they are as a set	—
16	DOES NOT LICENSE	That the ideas would work when carried out (see LD09-SL-C20)	—
17	VERIFICATION STATE	CONTENT VERIFIED PREPRINT	state: CONTENT_VERIFIED_PRIMARY · qualifiers: PREPRINT · verified_by: UNKNOWN
18	LIMITATIONS	One field; ideas judged before execution; the agent runs on an Anthropic model and this debate was drafted by an Anthropic model; read in arXiv version 1 (6 September 2024), which predates the ICLR 2025 acceptance	—
19	CONTRADICTORY / SUPPORTING SOURCES	Execution reverses the novelty advantage: LD09-SL-C20	—
20	GEOGRAPHY / LANGUAGE	Not checked / English	—
21	DISTRIBUTIONAL LEDGER	Not recorded	—
22	R01 / FIELD-MAP LINKS	LD-09, D3	—
23	WHAT WOULD CHANGE THIS CLAIM?	Replication in other fields and with other models	—
24	PUBLICATION STATUS	ELIGIBLE WITH SOURCE-POSITION QUALIFIER	state: ELIGIBLE · qualifiers: SOURCE_POSITION
25	RECORDED DATE	2026-10-15	—

Field 22 references resolved against Field Map v0.2: LD-09 (LIVING_DEBATE); D3 (DOMAIN)

The sources as recorded, with their record IDs and coded source types:

Pinned Sources (field 6, version-specific)

“Si, Yang & Hashimoto, “Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers,” ICLR 2025, arXiv:2409.04109” → Si et al. (2025), “Can LLMs Generate Novel Research…” (SRC-SL-SI-YANG-HASHIMOTO-ICLR-2025--v1.0) — Si, Yang & Hashimoto, “Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers,” ICLR 2025, arXiv:2409.04109; class PEER_REVIEWED_CONFERENCE_PAPER (explicit public provenance pin)

Cited by Living Debate versions

Individual gain, collective narrowing? Does AI make science, creative work and shared knowledge more alike? (v0.1) LD-09

Cite this page

SocioLens project. (2026, October 15). In a blind review by 79 NLP researchers, research ideas from a language-model agent were rated more novel than ideas from experts (5.64 against 4.84 on a 10-point scale) and slightly weaker on feasibility; but of 4,000 seed ideas the agent generated per topic, only about 200 were not duplicates [Claim LD09-SL-C19, version v1.0]. SocioLens. https://sociolens-web.pages.dev/claims/ld09-sl-c19/v1.0/

@misc{sociolens_ld09_sl_c19_v1_0,
  author = {{SocioLens project}},
  title = {In a blind review by 79 NLP researchers, research ideas from a language-model agent were rated more novel than ideas from experts (5.64 against 4.84 on a 10-point scale) and slightly weaker on feasibility; but of 4,000 seed ideas the agent generated per topic, only about 200 were not duplicates.},
  year = {2026},
  date = {2026-10-15},
  howpublished = {\url{https://sociolens-web.pages.dev/claims/ld09-sl-c19/v1.0/}},
  note = {Claim LD09-SL-C19, version v1.0}
}

The link points to this version, whose record never changes. For a finding, cite the original study as well: SocioLens summarizes research and does not replace it.

Research-of-record provenance

Materialized by sociolens-materializer 0.6.0 at 2026-09-28T20:00:00.000Z; normalization contract SocioLens — Claim Value Vocabulary v1.0.2 — Pre-Spike Materialization Contract (v1.0.2). Content seal SHA-256 f3e5daa91600d2927333b6773c57edfc3efd005399fd8dd90290b84878f45ae1.

Repository-authored inputs (not Drive exports)

SocioLens claim — LD09-SL-C19 v1.0 (v1.0) — written in the SocioLens repository by Project Lead (Claude, AI) on 2026-10-15; no Drive identity exists. Authority: AGENTS.md "Authority" (the Project Lead decides research content within the research-integrity rules) and the Owner's choice in the SocioLens project chat on 15 October 2026 to build the LD-09 debate; a new Claim, first version, written from the project's LD-09 source notes, checked against full texts or official abstracts. Block “ENTRY 19 — LD09-SL-C19” SHA-256 9c4d039da114d22fc6ab5d9181562f2d9a17052be36fcce1eed55b2793188926; fixture fixtures/authored/claims/LD09-SL-C19--v1.0.

Builder decisions (not Drive data)

VERSION = v1.0 — basis: Claim version taken from the repository-authored claim fixture version (SocioLens claim — LD09-SL-C19 v1.0); the first version of a new Claim of the LD-09 debate
CURRENT_POINTER = true — basis: repository-authored claim "SocioLens claim — LD09-SL-C19 v1.0" (Project Lead (Claude, AI), 2026-10-15); basis: AGENTS.md "Authority" (the Project Lead decides research content within the research-integrity rules) and the Owner's choice in the SocioLens project chat on 15 October 2026 to build the LD-09 debate; a new Claim, first version, written from the project's LD-09 source notes, checked against full texts or official abstracts
FIELD_6_SOURCE_REFERENCE = SRC-SL-SI-YANG-HASHIMOTO-ICLR-2025--v1.0 — basis: field 6 of ENTRY 19 — LD09-SL-C19 cites "Si, Yang & Hashimoto, “Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers,” ICLR 2025, arXiv:2409.04109"; the only field-6 segment of this ClaimVersion's repository-authored ledger block.
SOURCE_PROVENANCE_PIN = SRC-SL-SI-YANG-HASHIMOTO-ICLR-2025--v1.0 — basis: explicit version-pinned public provenance of this ClaimVersion (Amendment v1.2 §6): the pinned field-6 Source record "Si, Yang & Hashimoto, “Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers,” ICLR 2025, arXiv:2409.04109".