AI for Researchers · Visual Deck

Session 3: Reading, Notes & Knowledge Management

Grounded notebooks, structured extraction, and the nuance a summary quietly drops

{{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 · Foundations
the map · the taxonomy · CRIT"]):::done --> s2(["2 · Literature
the discovery matrix"]):::done s2 --> s3(["3 · Reading & notes
today"]):::today s3 --> s4(["4 · Writing
& integrity"]):::rest s4 --> s5(["5 · Data
& code"]):::rest s5 --> s6(["6 · Multi-
disciplinary"]):::rest s6 --> s7(["7 · Deep-
dives"]):::rest classDef done fill:#23204c,stroke:#55527e,color:#b9b7d6,font-size:18px classDef today fill:#ab7d22,stroke:#d9b36c,color:#14132b,font-size:19px classDef rest fill:#23204c,stroke:#8f8cb8,color:#8f8cb8,font-size:18px

Session 2's hard-won lesson: grounded ≠ correct — retrieval kills invented sources, not misread ones [11].
You now have a shortlist of papers. Today the shortlist becomes a corpus — and we find out what a summary of it leaves out.

AI for ResearchersLandscape as of August 20262 / 30

Learning Objectives

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


1Explain how paper Q&A / PDF-chat tools workand where summarization flattens methodological nuance
2Build a source-grounded notebook from a small set of papersGemini Notebook-style, formerly NotebookLM
3Run structured queries against a grounded notebookextract methods, effect sizes, and populations

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-paper comparison tablewith AI-assisted extraction
5Evaluate when a summary requires returning to the primary sourcerather than trusting the synthesis
6Integrate AI reading tools into a reference-manager workflowe.g., Zotero

Full verbatim wording in the reference deck and curriculum

AI for ResearchersLandscape as of August 20264 / 30

Section 01

What "Chat With This Paper" Actually Does

Before we trust an answer about a PDF, it is worth knowing which part of the PDF the tool was allowed to look at.

AI for ResearchersLandscape as of August 20265 / 30

Literacy Foundation

The Mechanism, in Four Steps


flowchart LR
  P["your PDF"]:::io --> A["1 · Index
broken into passages
the tool can look up"]:::n A --> B["2 · Retrieve
the passages that
look most relevant"]:::hot B --> C["3 · Generate
an answer from
those passages"]:::n C --> D["4 · Cite
the quoted passage —
click back to it"]:::n P -.-> L["2026 second path: million-token flagships
ingest a small corpus whole — no retrieval step"]:::ghost L -.-> C classDef io fill:#23204c,stroke:#8f8cb8,color:#eceaf8,font-size:19px classDef n fill:#23204c,stroke:#8f8cb8,color:#eceaf8,font-size:18px classDef hot fill:#1c1a3f,stroke:#d9b36c,color:#d9b36c,font-size:18px classDef ghost fill:#1c1a3f,stroke:#55527e,color:#8f8cb8,stroke-dasharray:6 6,font-size:16px

Google states it plainly: the notebook "retrieves the most relevant information based on your question first" [15] [16]. Step 2 is where most failures are born — and it is the step you cannot see.
The second path removes lost passages, not misreading — and at notebook scale, retrieval is still what happens.

AI for ResearchersLandscape as of August 20266 / 30

Tier 2 — Research-Specific Tools

Reading and Extraction Tools, From Their Own Documentation


Gemini Notebookformerly NotebookLM [14]
source-bound Q&A over a corpus you upload [15]; in-line citations — hover for the quote [16]
free tier: 50 sources per notebook, 50 chat queries/day [18]
Elicit
extraction tables — one column per data point [20]; click any cell for the supporting quote [21]
without the PDF it "can only pull information from the abstract" [19]
SciSpace
PDF chat plus a data extractor; "answers backed by citations from specific sections" [22]
maximum 50 extraction columns per table [23]
Paperpal
chat across uploaded sources, inside a writing workflow; "key findings with citations" [24]
writing-workflow scope; corpus limits not documented [24]
SCiNiTO
structured insights from an uploaded article, plus follow-ups; analysis history kept [25]
per-article analysis; cross-corpus use not documented [25]
Zotero + plugins
your library is the corpus; AI arrives as third-party code [26]; annotations carry page links [27]
"plugins have full access to your Zotero and your computer" [26]

Every cell from that tool's own current documentation — no cell is a comparative judgement. Row order is not a ranking; no endorsement implied.

AI for ResearchersLandscape as of August 20267 / 30

Critical Literacy

The Abstract Trap


the tool's own words"Only pull information from the abstract"Elicit's help centre, on what happens without the PDF [19]
why that mattersAbstracts are not neutral compressionsoutside the social sciences and humanities, "claims tended to be stronger in the abstract and conclusions than the discussion" [4]
the author's name for it"A weak hallucination""not stating a false fact but misrepresenting the strength of evidence for it" [4]

So the first question about any extraction table is: did the tool have the full text?

AI for ResearchersLandscape as of August 20268 / 30

Section 02

Where Summaries Flatten Nuance

Four independent lines of evidence, all pointing at the same thing: the qualifier is the first casualty.

AI for ResearchersLandscape as of August 20269 / 30

The Evidence — 1 of 4

Summaries Systematically Overgeneralise


26–73%of cases overgeneralised, for three of the ten models tested [1]
OR 1.90prompting for accuracy made overgeneralisation more likely [1]
OR 4.85vs. expert-written summaries of the same research [1]

4,900 summaries · 10 models of the 2025 cohort (ChatGPT-4o, Claude 3.7 Sonnet, DeepSeek, LLaMA 3.3 70B) [1] — commonest failure: "the patients in this trial" becomes "patients".
Within that cohort, "newer models tended to perform worse"; the 2026 follow-up finds the bias persists [33]. No study yet tests the August-2026 cohort — an open gap, not an all-clear.

AI for ResearchersLandscape as of August 202610 / 30

The Evidence — 2 of 4

Ask the Authors What Went Missing


first impression"Accurate but vague"the omissions surfaced only on a second reading — missing: context, disciplinary concepts, important findings and limitations, attribution to prior authors [2]
the tone shift"AI-summaries present author suggestions as established facts"verbatim from the study [2]
first casualtyMethods went firstthe section that carries "the validity of the findings" [2]

Twenty authors compared an AI summary of their own paper with their own [2] — qualitative, ChatGPT-4 (2025); cited for which content disappears, never as a rate.

AI for ResearchersLandscape as of August 202611 / 30

The Evidence — 3 of 4

Reading the Summary Instead Costs You Comprehension


Pre-registered randomised cross-over, n = 195, four GPT-4-era reading tools (2025) [3] — the authors: "much of the detail and nuance of the passage was lost in the summary".
Standardised test passages, not journal articles: take the direction; do not quote the effect size for research reading.

AI for ResearchersLandscape as of August 202612 / 30

The Evidence — 4 of 4

Even Experts Need a Checklist


20error patterns in the resulting schema — from fabricated citations to "synthesis failures in multi-document contexts" [5]
7categories those patterns fall into [5]
Unaidedexperts reliably caught errors of correctness and completeness [5]
Handed the schemathey caught "previously overlooked issues, particularly subtle hallucinations and citation errors" [5]

Scientists evaluating AI answers about papers they had written themselves [5]. A structured checklist beat unaided expert reading — that is why you are getting one.

AI for ResearchersLandscape as of August 202613 / 30

Series Artefact

The Summary-Trust Triage: When to Reopen the Paper


1Is there a number in it?effect sizes, n, CIs — go to the source. Always.
2Has a qualifier vanished?no population, no design, no boundary condition = a generic [1]
3Is it a limitation or a method?the two things summaries drop first [2]
4Does a claim carry no citation?unsupported prose in a grounded tool is a red flag, not a stylistic choice [16]
5Will you cite it?then the summary was never the evidence. The paper is [11]

Rule of thumb: use summaries to decide what to read and what to compare — never as the thing you cite.

AI for ResearchersLandscape as of August 202614 / 30

Section 03

Grounded Notebooks

A corpus you chose, a model that may only answer from it, and a citation on every sentence. What that buys — and what it does not.

AI for ResearchersLandscape as of August 202615 / 30

Grounded Notebooks

"Grounded" Is a Constraint, Not a Compliment


the defining rule"If the answer isn't in the source material, it won't provide a response."Google's own words [18]
the receiptsCitations are the quoted text itselfhover to read it, click to jump to it in the source — and you can tick sources in or out per question [16]
naming noteNotebookLM → Gemini Notebookrenamed 16 July 2026; AI-Pro tiers add a "secure cloud computer" for "data analysis grounded in your sources" [14]

The trade you are making: a smaller world, in exchange for being able to check every sentence. Vendor self-reported adoption, quoted as such: "more than 30 million people and over 600,000 organizations" [14] — nothing here depends on that number.

AI for ResearchersLandscape as of August 202616 / 30

Critical Literacy

Refusal Is a Capability — and It Is Unreliable


< 50%refusal accuracy on multi-document tasks, even for frontier models [6]
> 60%answerable questions refused, by some models at the other extreme [6]

RefusalBench: over 30 models of the 2024–25 cohort, six categories of flawed context, three intensity levels — "neither scale nor extended reasoning improves performance"; no re-test of the current generation yet [6]. Chat-with-paper benchmarks now score grounded refusal deliberately [7].
So: test your notebook's refusal before you trust its answers. We do exactly that in the demo.

AI for ResearchersLandscape as of August 202617 / 30

Grounded Notebooks

The Corpus Is the Method


1One question per notebooka mixed corpus produces mixed, unattributable answers
2Full texts, not abstractsthe whole point of a private corpus is the methods section [19]
3Record what you included and excludedany AI use that "makes or suggests judgements" must be reported [13]
50sources per notebook — the documented free-tier limit [18]
500,000words per source [18]
50chat queries a day (an upgrade is required to raise them) [18]

Two quiet gaps: copy-protected PDFs will not import, and footnotes and comments are not imported from Google files [17] [18]

AI for ResearchersLandscape as of August 202618 / 30

Section 04

From Reading to a Table

Twenty papers, one grid: methods, populations, effect sizes. The part of AI-assisted reading with the best evidence behind it — and the strictest conditions.

AI for ResearchersLandscape as of August 202619 / 30

Structured Extraction

Stop Asking for a Summary. Ask for a Column.


flowchart LR
  A["define columns
the fields you would
fill in by hand anyway"]:::n --> B["pilot
on three or four
papers first"]:::n B --> C["run
the same question of
every paper — rows are papers"]:::n C --> D["audit
click a cell,
read the supporting quote"]:::hot D --> E["export
CSV · Excel · RIS
into your own analysis"]:::n classDef n fill:#23204c,stroke:#8f8cb8,color:#eceaf8,font-size:18px classDef hot fill:#1c1a3f,stroke:#d9b36c,color:#d9b36c,font-size:18px

A summary is one blob of prose you cannot audit; a column is a defined question asked of every paper the same way [21] [23].
The same discipline as a systematic-review extraction form — the tool changes; the form does not.

AI for ResearchersLandscape as of August 202620 / 30

The Evidence

How Good Is It? It Depends Entirely on the Human


9.0%items incorrect, AI-assisted extraction [8]
11.0%items incorrect, human-only extraction [8]
41 minmedian time saved per study [8]

Six ongoing systematic reviews, blinded adjudicators, Claude 2.1→3.5 — a 2023–24 cohort — and a human checked every cell [8]. Take the human out: adjusted performance only "moderate" (GPT-4, 2024) [9].
Extraction stays Unreliable — the tradeoff persists: Opus 5 card (July 2026) reports accuracy up 11% and hallucinations up 6% [12] [34]; Cochrane, Campbell, JBI & CEE require human oversight [13]

AI for ResearchersLandscape as of August 202621 / 30

Series Artefact

The Extraction Column Template


NNameone phrase — "Primary outcome measured in the study"
IInstructionwhere to look, what counts, what format — give the unit ("mean ± SD"), or you will get five formats [20]
PPrecision"Be as precise as possible — '50 hours' is better than '2 days'." [20]
AAbstain rule"If the paper does not state this, write NOT REPORTED. Do not infer it."

Constrain the answer space where you can — Yes / No / Maybe, or a fixed list [20]. The abstain rule is the load-bearing line: without it, a blank becomes an inference.

AI for ResearchersLandscape as of August 202622 / 30

Today's Demo Corpus

Five Open-Access Papers, One Question: Does Open Access Raise Citations?


Davis et al. 2008BMJ 337, a568
Randomised controlled trial · 1,619 articles, 11 physiology journals
+89% full-text downloads; no citation advantage at one year [28]
Ottaviani 2016PLOS ONE 11(8)
Observational, matched control · 3,850 opened vs 89,895 closed
advantage "as high as 19%"; bigger for above-median articles [29]
Piwowar et al. 2018PeerJ 6, e4375
Large-scale observational · 3 samples × 100,000 articles
OA articles receive 18% more citations than average [30]
Langham-Putrow et al. 2021PLOS ONE 16(6)
Systematic review · 134 studies from 5,744 screened
47.8% found an advantage, 27.6% found none, 23.9% in subsets [31]
Huang et al. 2024Scientometrics 129
Large-scale observational · global outputs, 2010–2019
different outcome: OA is cited from more places [32]

All five open access with resolvable DOIs — re-verified against Crossref 2026-09-02; OA status via Unpaywall 2026-07-27. Full list in this session's demo script.

AI for ResearchersLandscape as of August 202623 / 30

Critical Literacy

The Column That Decides the Answer Is "Design"


18%what a summary will hand you, asked "does open access raise citations?" [30]
No citation advantagethe only randomised study in the corpus — and more readers [28]
A different outcomeone paper does not disagree at all — it measures where citations come from [32]

The review's sharpest finding is in its appraisal — of the three studies at low risk of bias: one advantage, one none, one in subsets [31].
A comparison table makes disagreement visible. A summary makes it disappear. [1]

AI for ResearchersLandscape as of August 202624 / 30

Section 05

Notes That Survive the Tool

Notebooks come and go. Your reference manager is where the reading has to land if you want it in two years' time.

AI for ResearchersLandscape as of August 202625 / 30

Knowledge Management

Where the AI Fits in a Zotero Workflow


What Zotero already does

  • highlights become notes that link back to the exact page [27]
  • "Show on Page" reopens the PDF where you made the annotation [27]
  • citations flow into Word, LibreOffice or Google Docs [27]

this is quote-to-source grounding — done by hand, and fully auditable

What the plugins add — and cost

  • AI arrives as third-party code, not a built-in feature [26]
  • Zotero's own warning: "plugins have full access to your Zotero and your computer" [26]
  • "an official plugin directory is planned" — there is still none [26]

lower-risk path: export the PDFs out to the reading tool instead [19]

AI for ResearchersLandscape as of August 202626 / 30

Systematic Prompting

CRIT, Applied to Reading


One paper, in PDF chat

Name the section, ask one thing, and demand the quote.

  • C which paper, which section
  • I quote the sentence you used; say NOT REPORTED if absent
  • T one field at a time — then re-ask

Many papers, in a notebook

Ask for a grid, per source, with disagreements kept intact.

  • C+R the corpus, the field, the register
  • I answer per source; never merge sources into one claim [5]
  • T then ask what the corpus cannot answer [6]
AI for ResearchersLandscape as of August 202627 / 30

Live Demo & Takeaways

What We Do Next — and What You Take Home


AWe build the five-paper notebook live, then ask it three questionsquestion 3 is one the corpus cannot answer — we watch whether it refuses, or obliges [6]
BThen: we extract the slide-23 comparison tableand check two cells against the PDFs
Take home: the Summary-Trust Triage · the Extraction Column Template · a notebook setup checklistone caution: AI reading tools reliably feel easier without reliably making you faster or more accurate — budget the verification time [10]
AI for ResearchersLandscape as of August 202628 / 30

References

References (1–16)


  1. Peters & Chin-Yee (2025). Generalization bias in large language model summarization of scientific research. R. Soc. Open Sci. 12(4), 241776.doi.org/10.1098/rsos.241776 — accessed 2026-08-21
  2. Hammond & Newell (2025). Reading between the GenAI lines. ASCILITE Publications (refereed poster).doi.org/10.65106/apubs.2025.2719 — accessed 2026-07-27
  3. Etkin, Etkin, Carter & Rolle (2025). Differential effects of GPT-based tools on comprehension of standardized passages. Frontiers in Education 10, 1506752.doi.org/10.3389/feduc.2025.1506752 — accessed 2026-07-27
  4. Thelwall (2026). Will AI be overconfident about academic research findings when reliant on abstracts? arXiv:2605.27392.arxiv.org/abs/2605.27392 — accessed 2026-07-27
  5. Martin, Humphreys, Brown, Leckey & Kaur (2026). An Expert Schema for Evaluating LLM Errors in Scholarly QA Systems. arXiv:2602.21059.arxiv.org/abs/2602.21059 — accessed 2026-07-27
  6. Muhamed, Ribeiro, Dreyer, Smith & Diab (2026). RefusalBench: Generative Evaluation of Selective Refusal in Grounded LMs. EACL 2026.aclanthology.org/2026.eacl-long.321.pdf — accessed 2026-07-27
  7. Imran & Solanky (2026). ResearchQA: Benchmarking Citation-Grounded QA on Scientific Papers. arXiv:2607.11074.arxiv.org/abs/2607.11074 — accessed 2026-07-27
  8. Gartlehner et al. (2025). AI-Assisted Data Extraction With a Large Language Model: A Study Within Reviews. Ann Intern Med 178(12).doi.org/10.7326/ANNALS-25-00739 — accessed 2026-07-27
  9. Khraisha, Put, Kappenberg, Warraitch & Hadfield (2024). Can LLMs replace humans in systematic reviews? Res. Synth. Methods 15(4).doi.org/10.1002/jrsm.1715 — accessed 2026-07-27
  10. Xu et al. (2026). Comparing structural and conversational AI summarisation for academic reading. Interact. Learn. Environ.doi.org/10.1080/10494820.2025.2604649 — accessed 2026-07-27
  11. Walters & Wilder (2023). Fabrication and errors in the bibliographic citations generated by ChatGPT. Sci. Rep. 13, 14045.nature.com/articles/s41598-023-41032-5 — accessed 2026-07-27
  12. Lieberum et al. (2025). LLMs for conducting systematic reviews: not yet ready for use. J. Clin. Epidemiol. 181, 111746.jclinepi.com/article/S0895-4356(25)00079-4 — accessed 2026-07-27
  13. Flemyng et al. (2025). Position statement on AI use in evidence synthesis (Cochrane, Campbell, JBI, CEE). Cochrane DSR, ED000178.cochranelibrary.com/cdsr/doi/10.1002/14651858.ED000178 — accessed 2026-07-27
  14. Google (2026). NotebookLM is now Gemini Notebook (announcement, 16 July 2026).blog.google/innovation-and-ai/products/gemini-notebook/notebooklm-gemini-notebook/ — accessed 2026-07-27
  15. Google (2026). Learn about Gemini Notebook. Gemini Notebook Help.support.google.com/gemininotebook/answer/16164461 — accessed 2026-07-27
  16. Google (2026). Use chat in Gemini Notebook. Gemini Notebook Help.support.google.com/gemininotebook/answer/16179559 — accessed 2026-07-27
AI for ResearchersLandscape as of August 202629 / 30

References

References (17–34)


  1. Google (2026). Add or discover new sources for your notebook. Gemini Notebook Help.support.google.com/gemininotebook/answer/16215270 — accessed 2026-07-27
  2. Google (2026). Frequently asked questions. Gemini Notebook Help.support.google.com/gemininotebook/answer/16269187 — accessed 2026-08-21
  3. Elicit (2026). Improve column results. Elicit Help Center.support.elicit.com/en/articles/14758163 — accessed 2026-07-27
  4. Elicit (2026). Create and save columns in Elicit. Elicit Help Center.support.elicit.com/en/articles/14758162 — accessed 2026-07-27
  5. Elicit (2026). Systematic Reviews in Elicit. Elicit Help Center.support.elicit.com/en/articles/14759154 — accessed 2026-07-27
  6. SciSpace (2026). Chat with any PDF | AI-powered Research Paper & PDF Summarizer.scispace.com/chat-pdf — accessed 2026-07-27
  7. SciSpace (2026). AI-Powered Data Extraction from Research PDFs.scispace.com/extract-data — accessed 2026-07-27
  8. Paperpal (2026). AI Academic Writing Tool — Comprehensive AI Research Assistant.paperpal.com/ — accessed 2026-07-27
  9. SCiNiTO (2026). PDF Analysis. SCiNiTO user guide.scinito.ai/help/pdf-analysis — accessed 2026-07-27
  10. Zotero (2026). Plugins for Zotero. Zotero Documentation.zotero.org/support/plugins — accessed 2026-07-27
  11. Zotero (2026). Zotero PDF Reader and Note Editor. Zotero Documentation.zotero.org/support/pdf_reader — accessed 2026-07-27
  12. Davis, Lewenstein, Simon, Booth & Connolly (2008). Open access publishing, article downloads, and citations: randomised controlled trial. BMJ 337, a568.doi.org/10.1136/bmj.a568 — accessed 2026-07-27
  13. Ottaviani (2016). The Post-Embargo Open Access Citation Advantage. PLOS ONE 11(8), e0159614.doi.org/10.1371/journal.pone.0159614 — accessed 2026-07-27
  14. Piwowar et al. (2018). The state of OA: a large-scale analysis of the prevalence and impact of Open Access articles. PeerJ 6, e4375.doi.org/10.7717/peerj.4375 — accessed 2026-07-27
  15. Langham-Putrow, Bakker & Riegelman (2021). Is the open access citation advantage real? A systematic review. PLOS ONE 16(6), e0253129.doi.org/10.1371/journal.pone.0253129 — accessed 2026-07-27
  16. Huang et al. (2024). Open access research outputs receive more diverse citations. Scientometrics 129(2), 825–845.doi.org/10.1007/s11192-023-04894-0 — accessed 2026-07-27
  17. Peters, Bertazzoli, DeJesus, van der Velden & Chin-Yee (2026). Generics in science communication: Misaligned interpretations across laypeople, scientists, and large language models. Public Underst. Sci. (online 20 Apr 2026).doi.org/10.1177/09636625261425891 — accessed 2026-08-21
  18. Anthropic (2026). System Card: Claude Opus 5 (24 July 2026; vendor self-evaluation).www-cdn.anthropic.com — accessed 2026-08-21