AI for Researchers · Visual Deck

Session 6: AI for Multidisciplinary Research

Getting oriented in a field that is not yours — in four rungs, with the checks that keep it honest

{{Presenter Name}}

Landscape as of August 2026

AI for ResearchersLandscape as of August 20261 / 30

2-Minute Recap

The Series So Far

flowchart LR
  s1(["1 · Map · Taxonomy
· CRIT"]):::past --> s2(["2 · Discovery
Matrix"]):::past --> s3(["3 · Summary-Trust
Triage"]):::past --> s4(["4 · Policy
Table"]):::past --> s5(["5 · Workflow Canvas
row 6 empty"]):::past --> s6(["6 · today
fill row 6"]):::today --> s7(["7 · deep-
dives"]):::rest
  classDef past fill:#23204c,stroke:#8f8cb8,color:#b9b7d6,font-size:17px
  classDef today fill:#ab7d22,stroke:#d9b36c,color:#14132b,font-size:18px
  classDef rest fill:#23204c,stroke:#55527e,color:#8f8cb8,font-size:17px

Today: the job no colleague has time to do for you — getting oriented in someone else's field,
fast enough to be useful and checked enough to be safe. Rule 2 rides along: rarity moves tasks rightward.

AI for ResearchersLandscape as of August 20262 / 30

Learning Objectives

By the End of This Session You Will Be Able To…

1Translate an unfamiliar discipline's jargon, methods and core literatureinto accessible terms
2Run a citation-graph or cross-field semantic searchadjacent literature, unexpected connections
3Evaluate evidence from fields with different methodological standardsfor compatibility, before synthesizing it

Objectives 4–6 next · full verbatim wording in the reference deck and curriculum

AI for ResearchersLandscape as of August 20263 / 30

Learning Objectives

…And You Will Also Be Able To

4Build a cross-disciplinary synthesis memocombining evidence from multiple fields
5Explain how AI can support team scienceshared knowledge bases · collaborator discovery · cross-expertise communication
6Produce an annotated reading listto orient into an unfamiliar field

Full verbatim wording in the reference deck and curriculum

AI for ResearchersLandscape as of August 20264 / 30

Section 01

Why Crossing Fields Is Hard

Not a confidence problem, and not a you problem. A measured, structural one — and knowing its exact shape tells you what to delegate.

AI for ResearchersLandscape as of August 20265 / 30

Literacy Foundation — Start With the Prize

The Highest-Impact Science: Mostly Conventional, One Foot Outside

Uzzi, Mukherjee, Stringer & Jones — Science 2013 · 17.9 million papers, all fields [1] A conventional core with one atypical intrusion wholly unusual work does not win — keep your question, your method and your standards; import one thing from elsewhere
2×"twice as likely to be highly cited" — papers pairing convention with a novel combination [1]
37.7%teams more likely than solo authors to insert novel combinations into familiar domains [1]

Funders ask for it by name — NSF convergence research: "deep integration across disciplines", a "shared scientific language" [15]. The one hard problem: finding the one thing, in a language you do not speak.

AI for ResearchersLandscape as of August 20266 / 30

The Evidence — Measured, Not Asserted

And It Is Priced In: The Bibliometric Penalties

18,476grant proposals: "the greater the degree of interdisciplinarity, the lower the probability of being funded" [6]
8 vs 20+years to half of a cohort ceasing to publish — most interdisciplinary vs moderately so, 154,021 biomedical PhDs [7]
128,950manuscripts incl. rejections: interdisciplinary topics lowered acceptance; interdisciplinary references raised it — cite across fields, frame within one [8]

The payoff is real but delayed: variety and disparity [12] lift long-term citations, depress short-term (3-year) ones [4]; both extremes lose [2] [3] — "a high-risk, high-reward endeavor" across ~900 scientists, 32,000 articles [5]. Indicators "should be interpreted with much caution" [13]: trust the direction, not the effect sizes.

AI for ResearchersLandscape as of August 20267 / 30

Critical Literacy — The Key Slide

The Mechanism Is Language. Then, Underneath It, Epistemology.

ways of knowingdisciplinary languages "reflect deeper differences in… epistemologies" — National Research Council, 2015 [11]
21,486articles: more jargon in title and abstract, fewer citations [9] — and the distance is field-pair-specific: ecology "balkanized by jargon", social sciences relatively integrated [10]

That 34% is about humans, not models — thirty minutes of searching makes you a non-expert with tabs open. This session is about the thirty minutes that does work.

AI for ResearchersLandscape as of August 20268 / 30

Section 02

AI as Field-Translator

Rung 1 of today's workflow — the one thing AI is unusually good at, and the one place it will mislead you most convincingly.

AI for ResearchersLandscape as of August 20269 / 30

Literacy Foundation

"Translate This Field" Is Three Different Requests

1Vocabulary — the twenty core termsand which also exist in your field with a different meaning — the half people skip. Close to Safe, only because you can check it
2Methods — default design, unit of analysis, what counts as positivewhere cross-field synthesis quietly goes wrong
3Canon — the five papers everyone citesnever from an ungrounded answer; grounded ≠ supported — Session 1's Dangerous zone, unchanged

Rule 1 — could you catch the error? — is the entry condition for all of today. A translation you cannot audit is a feeling of comprehension.

AI for ResearchersLandscape as of August 202610 / 30

The Terminology Trap — One Real Field Pair

False Friends: Same Word, Two Fields, Two Meanings

Clinical prediction researchMachine learning
DiscriminationDesirable — how well predictions "differentiate between individuals with and without the outcome" [57] [58]Undesirable — "unfair prejudice leading to inequity" [62] [59]
Bias"Systematic errors… from weaknesses in study designs and conduct" [63]"The intercepts in unit models" [63] — plus the fairness sense [59]. Three meanings, one word
ValidationExternal validation — so contested that TRIPOD+AI "refer[s] to validation as evaluation" [57]"Data used for parameter tuning" [57]; epi's external validation is ML's test set [63]
Calibration"Agreement between observed outcomes and estimated values" — routinely required [57] [58]Often absent: not addressed in 56 of 71 (79%) ML-vs-regression studies [60] — a missing concept

Every cell quoted from a guideline or peer-reviewed paper — never a model. Build this for your own field pair: rung 1's deliverable. Two more pairs (sensitivity = recall, PPV = precision [61]) in the reference deck.

AI for ResearchersLandscape as of August 202611 / 30

The Evidence — Two Results That Should Change How You Prompt

It Reads Better Than It Informs

Guo, Sohn, Leroy & Cohen — J. Biomedical Informatics 2026 · 150 participants, pre-2026 models [23] "Indistinguishable from human-written… [yet] human-written [summaries] lead to significantly better comprehension" comprehension and recall tested — clarity is the failure mode, not the goal
26–73%of cases overgeneralized, "even when explicitly prompted for accuracy" — prompting for accuracy roughly doubled it [22]
OR 4.85against human summaries — Peters & Chin-Yee 2025, 10 models of the 2024–25 cohort, 4,900 summaries; newer models did worse [22], and GPT-5-class models still overrate generalizability [64]

A real trade-off, documented across 53 teams: "between scoring highly for Factuality and… Relevance or Readability" [24]. The rung-1 check: can I find each definition in one field-B review article?

AI for ResearchersLandscape as of August 202612 / 30

Critical Literacy — Session 1's Rule 2, Quantified

It Is Weakest Exactly Where You Are Weakest

6% → 29%fabricated citations, well-covered vs niche topic — same model, same task (2025 study) [21]
39–77%factual support in the strongest grounded assistants — at link validity above 94% (May-2026 audit, 14 models) [66]
55.5%best tracked Humanity's Last Exam score (Claude Fable 5, Aug 2026) [65] — "low accuracy and calibration" at the expert frontier [19]
0.64best Macro-F1 at simply identifying interdisciplinary papers (random = 0.18; 2025 cohort) [26]

Counterfactual analogies — the core cross-field move: humans 75.3%, GPT-4 (2024) 45.2% [25]; benchmarks favour "vertical step-by-step reasoning typical of STEM" [20]. All of it is Rule 2: rarity moves tasks rightward.

AI for ResearchersLandscape as of August 202613 / 30

CRIT, Applied — Rung 1

The Translation Prompt, Built to Be Audited

C"A [FIELD A] researcher with no training in [FIELD B]"reading field-B papers well enough to judge whether their evidence bears on my question
RA field-B methodologist, explaining to a competent outsiderassume statistical literacy; assume no domain knowledge
I20 terms · flag every false friend, both meanings · mark "LOW CONFIDENCE"do not cite any paper, and do not use web search for this step
TA two-column glossary — then iteratethen verify every row against one field-B review article

"No papers, no web search" is deliberate: rung 1 is a vocabulary task, and citations are Session 1's Dangerous zone — [21] is why. Papers arrive on rung 2, from a database, or not at all.

AI for ResearchersLandscape as of August 202614 / 30

Section 03

Finding the Adjacent Literature

Rung 2 — where AI stops being the source and becomes the index. Three search geometries, and one question about who decided what field your paper is in.

AI for ResearchersLandscape as of August 202615 / 30

Literacy Foundation

Three Search Geometries — and Who Decides a Paper's Field

keywordMatches field B's wordsprecisely what you do not have yet — the one geometry that cannot work before rung 1
semanticMatches meaningOpenAlex embeds every title + abstract as a "1,024-dimensional vector" — returning "words your search never mentioned" [49]
citation graphMatches useConnected Papers: co-citation + bibliographic coupling — and "Connected Papers is not a citation tree" [43]

And the classification you browse is not neutral: WoS assigns "up to 6" categories at the journal level [52]; Scopus, "in-house experts" [51]; OpenAlex per work, automatically — 4 domains / 26 fields / ~252 subfields / ~4,500 topics [49]. Coverage gaps cluster "especially in the Humanities" [42].

AI for ResearchersLandscape as of August 202616 / 30

Series Artefact — Extending Session 2's Matrix

The Cross-Field Discovery Toolkit (1 of 2): Map the Landscape

Connected Papersone seed paper
similarity graph from co-citation + bibliographic coupling; "~50,000 papers" per graph [43]
"not a citation tree" — adjacency is similarity, not citation [43]
IncitefulLiterature Connector
two seeds: how two literatures connect through citations; 240M+ papers [44]
shortest paths only; one paper per end; max 6 hops [44]
ResearchRabbitcarried from Session 2
Similar / Earlier / Later work + author networks; 310M+ papers [45]
upstream corpus provider not documented; free tier capped at 50 seeds [45]
Litmapsmulti-seed
papers that "either cite or are cited by" the seed; 270M+ articles [46]
ranking behind "Discover" not documented [46]
Open Knowledge Mapsno seed paper needed
a query, clustered from word co-occurrence over titles, abstracts, keywords [47]
100 documents per map, by design — "a manageable amount" [47]
VOSviewerbring your own export
co-citation, coupling and term networks from a WoS / Scopus / PubMed export [48]
"free to use" — the site names no open-source licence, so nor does this deck [48]

Every cell from that tool's own documentation, fetched 2026-07-29. Where a vendor documents nothing, this deck says so. Not a ranking.

AI for ResearchersLandscape as of August 202617 / 30

Series Artefact — Extending Session 2's Matrix

The Cross-Field Discovery Toolkit (2 of 2): Search Sideways, Find People

OpenAlex semantic searchpaste a paragraph, not a phrase
1,024-dim vector over every title + abstract: "the richer the input, the better the matches" [49]
usage-priced since 2026 — "$1 of free usage every day"; 2,000-char input, 50 results, 1 req/s [49]
Semantic Scholar Recommendationsthe "unlike this" lever
positive and negative example papers — the documented way to push results away from your field; 214M papers [50]
pool is only "recent" or "all-cs"; the spec does not say it uses SPECTER2 [50]
OpenAlex topic filtersmove one level up
every work carries a primary_topic: 4 domains / 26 fields / ~252 subfields / ~4,500 topics [49]
the classification model is not documented — only "automatically assigned" [49]
OpenAlex Authors · ORCIDwho works in field B
author profiles filtered by topic, institution, country or ORCID [49] [56]
author queries are billable List+Filter calls [49] [56]
NIH RePORTERwho is funded right now
awards "from both NIH and non-NIH federal agencies" — earlier signal than papers [55]
US federal awards only; "no more than one URL request per second" [55]

Two recent changes catch people out: OpenAlex moved to usage-based pricing, and NotebookLM became Gemini Notebook (16 July 2026; re-verified 2026-08-21).

AI for ResearchersLandscape as of August 202618 / 30

Critical Literacy

Why Rung 2 Requires Two Tools

27 models"all models are less epistemically diverse than a basic web search" — larger models were less diverse; retrieval helped (through the 2025 generation; Aug-2026 cohort untested) [27]
3.02× · 4.84×AI-augmented scientists publish and are cited more — while topics studied shrink 4.63% and engagement between scientists falls 22% (41.3M papers, Nature) [16]

A similarity graph [43] and a shortest citation path [44] fail differently; two semantic tools fail the same way. AI makes you faster at crossing fields and science narrower — unless you use it to reach what it would hide.

AI for ResearchersLandscape as of August 202619 / 30

Section 04

Synthesis Across Evidence Standards

Rungs 3 and 4 — where "significant", "validated" and "replicated" turn out to name seven different things, and combining them naively is the failure mode.

AI for ResearchersLandscape as of August 202620 / 30

The Evidence

"A Significant Result" Names at Least Seven Different Bars

3 × 10⁻⁷particle physics: 5-sigma, "roughly a P value threshold" of this [29]
5 × 10⁻⁸genomics: the "genome-wide significance threshold" [29]
p < 0.05psychology, economics, social science — with 0.005 proposed for new discoveries [29]
a benchmarkCS / ML: a score on a shared benchmark — not a p-value at all; refereed conferences preferred [39]

Effects shrink too: half the size [31], 66% of original [32], "about 50%" [33], median 85% smaller [34]. Venues differ: arXiv 3,117,708 submissions since 1991; bioRxiv 37,648 preprints in five years [40]. The rows are not comparable to each other — that is the entire point.

AI for ResearchersLandscape as of August 202621 / 30

Series Artefact — Rung 4's Gate

The Cross-Field Evidence Compatibility Check

1What bar did this clear?"significant" spans p < 0.05 to 3 × 10⁻⁷ — and conclusions "should not be based only on whether a p-value passes a specific threshold" [29] [30]
2What is the unit of evidence, and who appraises it?medicine names its standards — GRADE [35] [36], RoB 2 [37], PRISMA 2020 [38]; ask field B to name the equivalent, and write down "we don't have one"
3What is the base rate of it holding up?36% · 61% · 62% · 46% — four fields, four differently measured figures [31] [32] [33] [34]
4Is the number even comparable?citation density differs by "about an order of magnitude" [41]; database coverage differs by field [42]
5Did you search where field B publishes?conferences, preprint servers and working papers are not a lesser tier everywhere [39] [40]

The rule: if you cannot answer all five, the finding enters your memo as a question for a colleague, never as evidence — what separates good cross-field work from poor is "how these are combined" [14].

AI for ResearchersLandscape as of August 202622 / 30

Section 05

Team Science, and the Workflow

The cheapest cross-field tool is a colleague. Here is what AI does for a mixed team — and the four-rung ladder that holds all of today together.

AI for ResearchersLandscape as of August 202623 / 30

Literacy Foundation

Team Science: AI Helps With Two of the Seven Problems

AI helps · 1 & 2Diversity & knowledge integrationthe named mechanism is language: "technical or scientific language unique to their area of expertise" [11] — a shared glossary, and a register-shifted methods section, are cheap and easy to check
a shared corpus helps · 5 & 6Boundaries & dispersiona Zotero group with "no limit on how many members may join" [53], or a grounded notebook over the team's own PDFs [54] — one corpus, one set of citations
AI does not help · 3, 4 & 7Size, goals, interdependencegovernance — no tool fixes them, and the National Research Council report does not suggest one does [11]

Collaborator discovery is a database query, not a chatbot question: author profiles by topic and institution [49] [56]; funded projects to see who is working on it now [55].

AI for ResearchersLandscape as of August 202624 / 30

Series Artefact · Today's Canonical Workflow

The Cross-Field Orientation Ladder

flowchart LR
  r1["1 · Translate
→ a glossary
─────
✓ every row vs one
field-B review article"]:::rung -->
  r2["2 · Map
→ a landscape
─────
✓ rebuilt in a structurally
different second tool"]:::rung -->
  r3["3 · Read
→ 8 annotated papers
─────
✓ all 8 resolved in a database
✓ 2 of 8 read in full yourself"]:::rung -->
  r4["4 · Synthesise
→ a one-page memo
─────
✓ Compatibility Check
✓ one field-B reader"]:::last
  classDef rung fill:#23204c,stroke:#d9b36c,color:#eceaf8,font-size:17px
  classDef last fill:#1c1a3f,stroke:#c14b58,color:#eceaf8,font-size:17px

You cannot skip a rung — each rung's output is the next rung's input, and each check makes that input safe. Tier 1 → Tier 2 → Tier 2 grounded → Tier 1; Unreliable throughout, Dangerous as evidence at rung 4. Why the checks: clarity is the failure mode [23] · one run is a sample [28] · valid links, 39–77% support [66] [21].

AI for ResearchersLandscape as of August 202625 / 30

Series Artefact — Session 5's Canvas, Row 6

Filling In the Row You Left Empty

1–5Frame · Discover · Read · Analyse · Write — filled last sessiongoverned by CRIT · Matrix · Triage · Guardrails · Policy Table — as recorded on your canvas
6Cross-field — today: field B's vocabulary, landscape and reading list — never its conclusionsglossary checked against one field-B review · every paper resolved · 2 of 8 read in full · one field-B reader on the memo — governed by the Cross-Field Orientation Ladder
7Your discipline's specialised toolsyou fill this row in during Session 7

Three rules: (1) the glossary is the deliverable, not the summary · (2) two tools, or no map [27] [28] · (3) the memo is a list of questions for a colleague — and when you publish, cite across fields, frame within one [8].

AI for ResearchersLandscape as of August 202626 / 30

Demo Preview — One Field Pair, Twenty-Five Minutes

A Clinical Epidemiologist Gets Oriented in Machine Learning

1Translate · 5 mintwenty terms, false friends flagged — then checked against a reporting guideline [57]; watch what it does with "validation"
2Map · 6 minone trusted seed into a similarity graph [43], then the same question into a shortest-citation-path [44] — two structurally different tools, on purpose [28]
3Read · 7 mineight papers, every DOI resolved live (expect all to resolve — no longer the test), then two opened and read against their annotation
4Synthesise · 5 mina memo from your notes, through the Compatibility Check — plus one deliberate failure: a finding that cannot legally cross

Everything runs on eight real, DOI-verified papers, six open access, listed in the handout — a pair chosen because discrimination, bias, validation and calibration mean different things on either side of it.

AI for ResearchersLandscape as of August 202627 / 30

References

References (1–21)

  1. Uzzi, Mukherjee, Stringer & Jones (2013). Atypical Combinations and Scientific Impact. Science 342.doi.org/10.1126/science.1240474 — accessed 2026-07-29
  2. Larivière & Gingras (2010). On the relationship between interdisciplinarity and scientific impact. JASIST 61(1).doi.org/10.1002/asi.21226 — accessed 2026-07-29
  3. Yegros-Yegros, Rafols & D'Este (2015). Does Interdisciplinary Research Lead to Higher Citation Impact? PLOS ONE 10(8).doi.org/10.1371/journal.pone.0135095 — accessed 2026-07-29
  4. Wang, Thijs & Glänzel (2015). Interdisciplinarity and Impact: Variety, Balance, Disparity. PLOS ONE 10(5).doi.org/10.1371/journal.pone.0127298 — accessed 2026-07-29
  5. Leahey, Beckman & Stanko (2017). Prominent but Less Productive. Administrative Science Quarterly 62(1).doi.org/10.1177/0001839216665364 — accessed 2026-07-29
  6. Bromham, Dinnage & Hua (2016). Interdisciplinary research has consistently lower funding success. Nature 534.doi.org/10.1038/nature18315 — accessed 2026-07-29
  7. Berkes, Marion, Milojević & Weinberg (2024). Slow convergence: Career impediments to interdisciplinary biomedical research. PNAS 121(32).doi.org/10.1073/pnas.2402646121 — accessed 2026-07-29
  8. Xiang, Romero & Teplitskiy (2025). Evaluating interdisciplinary research: Disparate outcomes for topic and knowledge base. PNAS 122(17).doi.org/10.1073/pnas.2409752122 — accessed 2026-07-29
  9. Martínez & Mammola (2021). Specialized terminology reduces the number of citations. Proc. R. Soc. B 288.doi.org/10.1098/rspb.2020.2581 — accessed 2026-07-29
  10. Vilhena et al. (2014). Finding Cultural Holes. Sociological Science 1.doi.org/10.15195/v1.a15 — accessed 2026-07-29
  11. National Research Council (2015). Enhancing the Effectiveness of Team Science. National Academies Press.doi.org/10.17226/19007 — accessed 2026-07-29
  12. Stirling (2007). A general framework for analysing diversity in science, technology and society. J. R. Soc. Interface 4(15).doi.org/10.1098/rsif.2007.0213 — accessed 2026-07-29
  13. Wang & Schneider (2020). Consistency and validity of interdisciplinarity measures. Quantitative Science Studies 1(1).doi.org/10.1162/qss_a_00011 — accessed 2026-07-29
  14. McLeish & Strang (2016). Evaluating interdisciplinary research. Palgrave Comms 2, 16055.doi.org/10.1057/palcomms.2016.55 — accessed 2026-07-29
  15. U.S. National Science Foundation (2026). Learn About Convergence Research.nsf.gov/funding/learn/research-types — accessed 2026-07-29
  16. Hao, Xu, Li & Evans (2026). AI tools expand scientists' impact but contract science's focus. Nature 649.doi.org/10.1038/s41586-025-09922-y — accessed 2026-07-29
  17. Rein et al. (2024). GPQA: A Graduate-Level Google-Proof Q&A Benchmark. COLM 2024.arxiv.org/abs/2311.12022 — accessed 2026-07-29
  18. Wang, Ma, Zhang et al. (2024). MMLU-Pro. NeurIPS 2024 Datasets & Benchmarks.arxiv.org/abs/2406.01574 — accessed 2026-07-29
  19. Phan et al. (2026). A benchmark of expert-level academic questions to assess AI capabilities (HLE). Nature 649.doi.org/10.1038/s41586-025-09962-4 — accessed 2026-07-29
  20. Kang et al. (2026). HSSBench: Humanities and Social Sciences Ability for Multimodal LLMs. Preprint.arxiv.org/abs/2506.03922 — accessed 2026-07-29
  21. Linardon et al. (2025). Influence of Topic Familiarity and Prompt Specificity on Citation Fabrication. JMIR Mental Health 12.doi.org/10.2196/80371 — accessed 2026-07-29
AI for ResearchersLandscape as of August 202628 / 30

References

References (22–41)

  1. Peters & Chin-Yee (2025). Generalization bias in large language model summarization of scientific research. R. Soc. Open Sci. 12(4).doi.org/10.1098/rsos.241776 — accessed 2026-07-29
  2. Guo, Sohn, Leroy & Cohen (2026). Are LLM-generated plain language summaries truly understandable? J. Biomed. Inform. 179.doi.org/10.1016/j.jbi.2026.105038 — accessed 2026-07-29
  3. Goldsack, Scarton, Shardlow & Lin (2024). Overview of the BioLaySumm 2024 Shared Task. BioNLP @ ACL 2024.aclanthology.org/2024.bionlp-1.10/ — accessed 2026-07-29
  4. Lewis & Mitchell (2024). Using Counterfactual Tasks to Evaluate the Generality of Analogical Reasoning in LLMs. Preprint.arxiv.org/abs/2402.08955 — accessed 2026-07-29
  5. Shen et al. (2026). IDRBench: Understanding the Capability of LLMs on Interdisciplinary Research. Preprint.arxiv.org/abs/2507.15736 — accessed 2026-07-29
  6. Wright et al. (2026). Epistemic Diversity and Knowledge Collapse in Large Language Models. Preprint.arxiv.org/abs/2510.04226 — accessed 2026-07-29
  7. Dathe, Hoffmann & Mangold (2026). Useful for Exploration, Risky for Precision: Evaluating AI Tools in Academic Research. Preprint.arxiv.org/abs/2605.10125 — accessed 2026-07-29
  8. Benjamin et al. (2018). Redefine statistical significance. Nature Human Behaviour 2(1).doi.org/10.1038/s41562-017-0189-z — accessed 2026-07-29
  9. Wasserstein & Lazar (2016). The ASA Statement on p-Values. The American Statistician 70(2).doi.org/10.1080/00031305.2016.1154108 — accessed 2026-07-29
  10. Open Science Collaboration (2015). Estimating the reproducibility of psychological science. Science 349.doi.org/10.1126/science.aac4716 — accessed 2026-07-29
  11. Camerer et al. (2016). Evaluating replicability of laboratory experiments in economics. Science 351.doi.org/10.1126/science.aaf0918 — accessed 2026-07-29
  12. Camerer et al. (2018). Evaluating the replicability of social science experiments in Nature and Science. Nature Human Behaviour 2.doi.org/10.1038/s41562-018-0399-z — accessed 2026-07-29
  13. Errington et al. (2021). Investigating the replicability of preclinical cancer biology. eLife 10.doi.org/10.7554/eLife.71601 — accessed 2026-07-29
  14. GRADE Working Group — Schünemann et al., Eds. (2013). GRADE Handbook.gdt.gradepro.org/app/handbook — accessed 2026-07-29
  15. Balshem et al. (2011). GRADE guidelines: 3. Rating the quality of evidence. J. Clin. Epidemiol. 64(4).doi.org/10.1016/j.jclinepi.2010.07.015 — accessed 2026-07-29
  16. Sterne et al. (2019). RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 366.doi.org/10.1136/bmj.l4898 — accessed 2026-07-29
  17. Page et al. (2021). The PRISMA 2020 statement. BMJ 372.doi.org/10.1136/bmj.n71 — accessed 2026-07-29
  18. Patterson, Snyder & Ullman (1999). Evaluating Computer Scientists and Engineers For Promotion and Tenure. CRA Best Practices Memo.cra.org/resources/best-practice-memos — accessed 2026-07-29
  19. arXiv (2026). Monthly submission statistics; and Abdill & Blekhman (2019). Tracking the popularity and outcomes of all bioRxiv preprints. eLife 8.arxiv.org/stats/monthly_submissions · doi.org/10.7554/eLife.45133 — accessed 2026-07-29
  20. Waltman (2016). A review of the literature on citation impact indicators. J. Informetrics 10(2).doi.org/10.1016/j.joi.2016.02.007 — accessed 2026-07-29
AI for ResearchersLandscape as of August 202629 / 30

References

References (42–66)

  1. Martín-Martín et al. (2021). A multidisciplinary comparison of coverage via citations. Scientometrics 126(1).doi.org/10.1007/s11192-020-03690-4 — accessed 2026-07-29
  2. Connected Papers (2026). About; Pricing.connectedpapers.com/about · /pricing — accessed 2026-07-29
  3. Inciteful (2026). Paper Discovery; Literature Connector; Data Sources.incitefulmed.com/academic/ — accessed 2026-07-29
  4. ResearchRabbit (2026). Home; Pricing; Help guide.researchrabbit.ai — accessed 2026-07-29
  5. Litmaps (2026). Pricing; Documentation.litmaps.com/pricing · docs.litmaps.com — accessed 2026-07-29
  6. Open Knowledge Maps (2026). About; FAQ.openknowledgemaps.org — accessed 2026-07-29
  7. VOSviewer (2026). Download; Features. v1.6.21, 12 June 2026.vosviewer.com — accessed 2026-07-29
  8. OpenAlex (2026). Topics; Semantic search; Authors; Authentication & pricing.developers.openalex.org · api.openalex.org — accessed 2026-07-29
  9. Semantic Scholar (2026). Academic Graph API; Recommendations API.api.semanticscholar.org — accessed 2026-07-29
  10. Elsevier (2024/2026). What is the complete list of ASJC subject areas in Scopus?service.elsevier.com — accessed 2026-07-29
  11. Clarivate (2025). Web of Science Core Collection: Web of Science Categories.support.clarivate.com — accessed 2026-07-29
  12. Zotero (2026). Groups; Storage.zotero.org/support/groups — accessed 2026-07-29
  13. Google (2026). Gemini Notebook (formerly NotebookLM) — FAQ; plans.support.google.com/gemininotebook — accessed 2026-07-29
  14. NIH (2026). NIH RePORTER APIs.api.reporter.nih.gov — accessed 2026-07-29
  15. ORCID (2026). What is ORCID?info.orcid.org — accessed 2026-07-29
  16. Collins et al. (2024). TRIPOD+AI statement: updated guidance for reporting clinical prediction models. BMJ 385.doi.org/10.1136/bmj-2023-078378 — accessed 2026-07-29
  17. Van Calster et al. (2019). Calibration: the Achilles heel of predictive analytics. BMC Medicine 17.doi.org/10.1186/s12916-019-1466-7 — accessed 2026-07-29
  18. Obermeyer et al. (2019). Dissecting racial bias in an algorithm used to manage the health of populations. Science 366.doi.org/10.1126/science.aax2342 — accessed 2026-07-29
  19. Christodoulou et al. (2019). No performance benefit of machine learning over logistic regression. J. Clin. Epidemiol. 110.doi.org/10.1016/j.jclinepi.2019.02.004 — accessed 2026-07-29
  20. Assel & Vickers (2025). The F score ranks diagnostic tests and prediction models inconsistently with their clinical utility. Diagn. Progn. Res. 9(1).doi.org/10.1186/s41512-025-00214-7 — accessed 2026-07-29
  21. Paulus & Kent (2020). Predictably unequal: algorithmic clinical prediction may increase health disparities. npj Digital Medicine 3.doi.org/10.1038/s41746-020-0304-9 — accessed 2026-07-29
  22. Sung & Hopper (2023). Co-evolution of epidemiology and artificial intelligence. Int. J. Epidemiol. 52(4).doi.org/10.1093/ije/dyad089 — accessed 2026-07-29
  23. Peters et al. (2026). Generics in science communication. Public Underst. Sci., online 20 Apr 2026.doi.org/10.1177/09636625261425891 — accessed 2026-08-21
  24. Artificial Analysis (2026). MMLU-Pro and Humanity's Last Exam evaluation pages (independent runs).artificialanalysis.ai/evaluations — accessed 2026-08-21
  25. Onweller et al. (2026). Cited but Not Verified: Parsing and Evaluating Source Attribution in LLM Deep Research Agents. Preprint.arxiv.org/abs/2605.06635 — accessed 2026-08-21
1 / 30