Video summary

WEBINAR SPESIAL Penyusunan Artikel dengan Asistensi JadimajuAI

Main summary

Key takeaways

Technology

Webinar focus: Writing research articles with “JadiMajuAI” assistance

The webinar explains how to use multiple AI tools and free/paid bibliometric workflows to quickly find the state of the art, identify research gaps, formulate novelty, and then move toward drafting (and later validating) an academic manuscript.

1) AI tool selection (purpose-based)

Speakers emphasize that different AI tools serve different needs:

  • Gemini (e.g., Gemini 2.5 Pro): best for brainstorming, discussion, and research ideation (e.g., improving topic direction, exploring angles, brainstorming keyword sets).

  • Perplexity: best for searching and collecting statistical/research results.

  • Dify (mentioned as “dipsych”): for deeper analysis.
  • ChatGPT (referred to as “5” in subtitles): best for writing (not primarily for reference discovery).

Also noted: some tasks (e.g., route planning / transport ordering like Grab) are not where AI is best—use specialized services instead.

2) Choosing a research topic using gaps & novelty (core method)

The webinar frames article quality as:

  • State of the art → overview of existing research
  • Gap research → identifying what has not been studied yet (or lacks evidence)
  • Novelty → the “new contribution” derived from the gap

Key message: AI cannot judge novelty/gaps well unless you provide enough data and references (and/or use tools that already index scholarly content).

3) Two ways to identify gaps: bibliometrics vs literature review

A. Bibliometrics (fast; works at scale)

  • Uses metadata fields such as: publication year, author, journal, keywords, citations.
  • Claimed advantage: can process thousands of papers in minutes (example given: ~15 minutes for an initial overview).
  • Limitation: bibliometrics can be superficial (doesn’t capture rich contextual detail).

B. Literature review (slower but deeper)

  • Requires reading and synthesizing the actual text: introduction, methodology, results, discussion.
  • Advantage: captures context and nuance (example contrasted “simple quantification” vs deep lived constraints and budgeting logic).

Recommended approach

Use both:

  • Bibliometrics to map gaps quickly
  • Literature review to write with deeper justification

Novelty can be built in different “weight levels” (and the webinar stresses that novelty is subjective, depending on standards and discipline).

4) Examples of novelty types (time, location, object/theory/context)

Novelty is illustrated as not always “heavy”:

  • Time novelty: same topic repeated at a different time (e.g., leadership style re-studied across different years).

  • Location novelty: same general theory/topic studied in a new place/context (e.g., regency vs city; different community settings).

  • Theory/contextual novelty: using different theories/models to reinterpret a broad topic.

  • Important note: acceptance of novelty (especially for dissertation vs thesis) depends on required depth—dissertations often require theory + methodology + contextualization.

5) Free bibliometric workflow tools + why English keywords

Why English keywords?

Because much internationally indexed literature uses English, enabling better coverage in databases (compared with Indonesian-only journal categorization such as Sinta).

Tools mentioned for free access

  • Google Scholar (noted, but tools may get blocked during automated use).
  • Crossref (recommended for general journal metadata coverage; speaker claims indexed journals can be queried widely).

Workflow using “Publish or Perish” style tooling + “Publish/PoP viewer”

Steps demonstrated conceptually:

  1. Prepare 3 English keywords.
  2. Run bibliometric queries (e.g., within a year range like 2021–2025).
  3. Export results as RIS.
  4. Load into a visualization tool (“Post Viewer” mentioned) to generate:
    • Network visualization (clusters/connected themes)
    • Overlay visualization (trend over years)
    • Density visualization (research intensity/overcrowding)

6) Visual analytics for gap & novelty justification

Bibliometric visuals are treated as evidence to justify:

  • State of the art maturity: dense, connected clusters indicate well-developed areas.
  • Gap identification, such as:
    • Missing keywords/concepts (e.g., “coastal communities” not appearing strongly)
    • Weak or absent connections between concepts (no “link lines” between clusters)
    • Context gaps (a topic exists but not in the intended neglected context)
  • Novelty framing, such as:
    • Emphasizing context change (e.g., coastal communities + inclusive education access + disability)
    • Highlighting lack of strong empirical linkage between social capital and access/equity in that specific context

Example justification described in subtitles:

  • State of the art: education/postgraduate research area is “ripe/mature” (busy clusters)
  • Gap: contextual gap—coastal communities are not prominent
  • Novelty: contextual and integrative gap—equity in access for children with disabilities in coastal contexts not strongly connected in prior work; claim of first empirical evidence

7) Practical manuscript writing flow with AI (prompting + formatting)

After justification:

  1. Use AI to draft:

    • Title (guided by an object–function–method format; max 15–20 words)
    • Outline / literature mapping (3 main points; each expected to yield ~8 references)

  2. Use AI tools for section drafting with structured prompts:

    • Introduction / background / problem formulation (~1000 words in Indonesian; each paragraph minimum sentences)
    • Methods (research design, subjects/sampling, instruments, procedures, analysis; detailed for replicability)
    • Results (present data without interpretation; include tables/graphs/diagrams; interview excerpts mentioned)
    • Discussion (connect findings to the provided references only; explain significance, implications, limitations)
    • Conclusion (~300 words)
    • Abstract (150–250 words; objectives, methods, main results, conclusion; include keywords)

Speaker’s repeated notes:

  • Use only the references you provide (don’t add new citations).
  • Adjust output structure (e.g., sentence minimums, numbering).

8) Transcription for qualitative data + contextual interpretation

For qualitative research, the webinar includes:

  • Creating instruments from theory + problem formulation (example: Putnam social capital categories).
  • Transcribing interviews using TurboScribe (named “Turbo Scrap” in subtitles):
    • Free plan limits (example: up to 3 transcripts/day, each up to 30 minutes total recording)
    • Paid plan offers longer/unlimited transcription (speaker mentions “unlimited,” with caveats)

Crucial: interpretation matters after transcription. Example: “astagfirullahalazim” can mean different things depending on conversational context; AI transcription alone may misread intent without researcher interpretation.

9) Citations, exporting references, and bibliography management

Tools/process mentioned:

  • Export references as RIS from bibliometric tool queries.
  • Use Mendeley Desktop plus Mendeley Web Importer.
  • Rapid Journal Quality Check for identifying journal status (e.g., Q1/Q2/Q3).
  • Use Mendeley to insert citations into Word automatically (replace date/author-year style).

10) Manuscript quality checks (AI detection, paraphrasing, translating)

The webinar describes a multi-step “submission readiness” flow:

  • AI-detector test (example: using GPTZero; speaker suggests premium tools for detection).
  • Paraphrasing to reduce AI-detection risk (via large rewrites; subtitles contain garbled phrasing but the intent is to paraphrase substantial portions).
  • Translate Indonesian → English using a Word/AI translation feature.
  • Check journal targeting:
    • Rapid Journal Quality Check + Google Scholar title checks
    • Verify publisher/journal template and APC guidance
  • Insert citations and ensure reference list integrity.

Plagiarism stance/caution:

  • Plagiarism is determined by similarity testing (e.g., Turnitin workflow mentioned conceptually).
  • AI will not “guarantee” no plagiarism—still must pass similarity/forensic/ethics checks.
  • Speaker claims AI-generated summaries can reduce similarity, but the core requirement remains: correct citation and genuine originality.

11) Account guidance & warnings (sharing vs personal)

The webinar emphasizes:

  • Prefer a personal account, not sharing, to avoid:
    • slower performance/timeouts
    • export/RIS errors
    • potential bias or data exposure risk (“risk of data leaks”)
  • Speaker attributes tool failures to shared account usage and high request volume.

Main speakers / sources (as inferred from subtitles)

  • Khairul Anam (Jombang, East Java) — primary presenter for the technical workflow and AI/bibliometrics guidance.
  • Mas Pawan / Prof. Wawan — host/moderator introducing topics and coordinating the session.
  • Jadi Maju Indonesia team — referenced as the organizer/host channel.

Original video