Who Owns the Narrative? �Data, Disinformation, and the Missing Voice of Academia

Executive Summary — Who Owns the Narrative? Data, Disinformation, and the Missing Voice of Academia

Purpose. This keynote argues that social media is today’s battlefield of narratives. If scholars stay on the sidelines, public discourse is ceded to buzzers, bots, and misinformation. The talk outlines a practical path for academics to lead data-driven public literacy at scale.

Evidence base. Drawing on Drone Emprit’s decade of monitoring billions of social and news conversations, the talk shows how coordinated behavior and bot-like activity shape “buzz,” and contrasts U.S. universities—where academics act as influential public communicators—with Indonesian universities, where engagement is largely institutional and reactive.

Key findings.

Bots vs. humans: Coordinated accounts can inflate attention and tilt frames unless countered by credible expert voices.

Participation gap: U.S. networks (e.g., MIT/Stanford/Harvard) feature dense clusters of academics and alumni; Indonesian networks center on official accounts with sparse scholar participation—weakening research-based contributions to public debates.

Barriers to entry: Political pressure, legal/reputational risks, workload, and limited public-communication skills discourage academics from speaking up.

Practical model for safe engagement.

Constructive journalism lens: Focus on data, pair critique with solutions, and use accessible language; avoid partisan labeling and personal attacks.

AI as enabler: Any LLM (ChatGPT, Grok, Gemini, Claude, etc.) can help scholars distill research into plain, platform-ready outputs—threads, LinkedIn posts, short video scripts—while maintaining citations and institutional ethics.

Crisis framework example: Applying SCCT to a real case (Free Nutritious Meals/MBG) shows how to classify crisis type, read public attribution, and craft a “Rebuild” response strategy rooted in transparency, corrective action, and empathy.

Recommendations.

Normalize scholar presence online (policies, training, and recognition for public-facing scholarship).

Adopt a “data-first, solutions-oriented” style anchored in recognized theories (SCCT, Entman framing, stakeholder mapping).

Operationalize AI workflows for fast, accurate, and ethical public communication (brief → distill → localize → format → verify).

Measure impact beyond likes: track clarity signals (saves, meaningful comments), media citations, and policy uptake.

Call to action. Don’t stop at journals. Become visible, constructive, data-driven voices. If academics don’t lead the narrative with evidence, others will.

1. Who Owns the
Narrative?
Data, Disinformation, and the
Missing Voice of Academia
Ismail Fahmi, Ph.D.
Director of Media Kernels Indonesia (Drone Emprit)
[email protected]

2. 2
1992 – 1997 Undergraduate, Electrical Engineering, ITB, Indonesia
2003 – 2004 Master, Information Science, University of Groningen, NL
2004 – 2009 Doctor, Information Science, University of Groningen, NL
2009 – Now Consultant at Weborama (Paris/Amsterdam)
2014 – Now Founder PT. Media Kernels Indonesia, a Drone Emprit Company
2017 – Now Lecturer at the Magister Program of the Universitas Islam Indonesia
2021 – Now Vice Chairman at the Infocomm Commission, Majelis Ulama Indonesia
2023 – Now Vice Chairman at the MPI, Central Board of Muhammadiyah
2024 – Now Founder/Preskom at the LabMu (Muhammadiyah Software Labs)
Ismail Fahmi, Ph.D.
[email protected]
Bojonegoro, 1974

3. Background:
The Story of a Narrative War
3
Social Network Analysis “RUU KUHP” – Drone Emprit (2019)
“Hashtags are not
just symbols. They
are digital battle
flags. When
thousands rally
behind them, a
war of opinion
erupts: whoever
controls the
narrative, controls
perception.”

4. The Big Question
4
If social media has become the new battlefield of
narratives,
Shouldn’t academics be on the frontlines --
educating the public with data,
Not leaving the field to buzzers and bots?

5. Drone Emprit: A Living Lab
5
Since 2014:
monitoring
billions of
conversations
from social
media & online
news

6. Data-Driven Public Literacy
6
Applied in:
elections,
COVID-19
hoaxes, social
conflicts, public
policy debates

7. From Data to Knowledge
The Challenges
7
BIG DATA ≠ AUTOMATIC
KNOWLEDGE
CHALLENGES:
NOISE, MANIPULATION,
BOTS
LINKED DATA:
CONNECTING SOCIAL +
HEALTH + ECONOMY FOR
HOLISTIC INSIGHTS

8. Behind the Buzz: How Bots and Humans Compete in Storytelling
8

9. When Patterns Expose the Hidden Hands

10. Academics on Social Media
USA VS Indonesia
10

11. Social Network Analysis: @mit, @stanford, @harvard
11

12. Direct Network of @Harvard
12
Key Insight – Harvard
Network
•Harvard’s direct
network shows high
engagement of
academics and alumni in
public conversations,
strongly supported by
media and influencers.
•This strengthens
academic-based public
literacy, since many
expert voices actively
contribute to the flow
of information.

13. SNA @itbofficial, @UGMYogyakarta, @univ_indonesia
13
OppositionPro Government

14. Direct Network of @itbofficial
14
Brian Yuliarto
(Now Mendiktisaintek)
Key Insight – ITB Network
•@itbofficial functions
more as a digital notice
board rather than a
space for public debate
and knowledge
exchange.
•Academic and alumni
engagement is minimal,
making the university’s
voice in public
conversations almost
unheard.

15. Topics of Conversation for Harvard vs ITB
15
Harvard’s discourse is global, political,
and activist-driven. Academics and
students use Twitter to engage with
pressing international issues, advocacy
campaigns, and reputational debates.
Harvard is not just a university brand, but
an actor in global conversations.
ITB’s discourse is institutional, local,
and commemorative. Most conversations
focus on campus events, obituaries,
student movements, and Bandung-related
topics. Less involvement in global activism
or international academic debates.

16. Findings: US v.s. Indonesian Universities
US Universities (MIT, Stanford, Harvard)
•Dense Academic ParticipationThe ellipses highlight many individual accounts of professors, researchers, experts, and technopreneurs. They are not only present, but also highly connected to the central clusters of their universities.
•Academics as InfluencersAcademics appear as influential nodes, actively retweeted and mentioned. They help amplify institutional voice while also sharing their independent perspectives.
•Conversation TopicsInvolve science communication, technology, innovation, policy debates, and academic insights.
•Impact: Universities are knowledge hubs on social media, with academics actively shaping public discourse and extending their role beyond journals.
Indonesian Universities (ITB, UGM, UI)
•Institutional DominanceThe central nodes are mainly the official university accounts (@itbofficial, @UGMYogyakarta, @univ_indonesia).
•Sparse Academic InvolvementFew ellipses (academics/experts) are visible, and they are much less central in the networks. Most conversations cluster around political alignments (pro-government vs opposition) rather than around individual academics.
•Academics as Passive NodesUnlike the US case, Indonesian academics rarely appear as independent influencers who actively engage with the public.
•Impact: Universities are seen more as institutions, not as communities of engaged scholars on social media.
16

17. Comparative Insights
17
US Model: Academics are frontline communicators. They tweet, debate, and
educate the public directly → building trust and credibility.
Indonesian Model: Academics are mostly silent observers. Public discourse
is dominated by institutional accounts and political clusters, leaving little
space for data-driven or research-based input from scholars.
Implication: This gap reduces the role of Indonesian academia in shaping
public literacy and national discourse.

18. Drone Emprit Academic (DEA)
Mission: Data-driven Public Literacy
18
2018–2023: free access to Twitter/X data and analytics for academics, researchers, journalists
Produced hundreds of publications: theses, dissertations, journals, conference papers
Proved that open data accelerates academic research

19. Drone Emprit AcademicExpectation vs Reality
19
Expectation: academics
use DEA → real-time
analysis → inform &
educate the public
Reality: most works
stayed in journals; very
few academics engaged
the public with data
“If academics remain
absent, public debates will
keep losing to buzzers and
disinformation.”

20. Why Many Academics Stay Silent on Social Media?
Political Pressure & Academic Freedom
•Fear of state repression and institutional constraints.
•Sensitive issues often discouraged in universities.
Spiral of Silence
•Fear of backlash, isolation, or negative labeling.
•Safer to stay quiet than risk being targeted.
Legal & Reputational Risks
•Potential consequences under ITE Law, defamation suits, or internal sanctions.
Professional Workload
•Research, teaching, and admin duties leave little energy for public engagement.
Weak Public Communication Skills
•Comfortable in journals & conferences, less skilled at simplifying knowledge for wider audiences.
Closed Academic Culture
•Tradition of “neutrality” → academics avoid being seen as political actors.
20

21. Safe Strategies for Academics on Social Media
Learning from Constructive Journalism
21
Focus on Data, Not PoliticsPresent evidence & analysis (charts, sentiment, SNA).
Avoid personal attacks or partisan framing.
Frame Criticism as Constructive
Highlight problems and offer solutions.
Example: Instead of “government failed,” say “data
shows improvement is needed in monitoring kitchens.”
Use the Language of Public Education
Simplify theory (e.g., SCCT, Entman) into easy messages for citizens.
Narratives: “This is not an accident, it’s preventable.
Here’s how we fix it.”
Adopt Constructive Journalism
Approach
Report problems, but also pathways forward.
Balance: critique + empathy + solution = safer and
more impactful.

22. Constructive Journalism
Core Principles
22
Problem + Solution
•Don’t just
highlight
problems →
always include
possible solutions.
•Ex: “Data shows
unsafe kitchens
→ Audit +
Certification
required.”
Hope & Agency
•Show that change
is possible and
actions matter.
•Ex: “MBG can
succeed if
transparency &
safety are
prioritized.”
Data-Driven
Storytelling
•Base narratives
on data, not
personal opinions.
•Use visuals,
charts, or simple
analogies to
educate the
public.

23. From Silence to Influence
AI as an Enabler for Academics
23

24. AI Specialization
Mastering Theories, Frameworks, and Disciplines
24

25. Media Analysis with Theoretical Models
Examples of Popular Frameworks
25
Entman → framing of
political issues
SCCT → situational crisis
communication theory
Van Dijk → discourse
analysis of public
programs
Stakeholder Mapping →
actors: buzzers, media,
politicians

26. Applying “SCCT” with AI for Crisis Analysis:
The MBG Food Poisoning Case
26

27. Big Data Collection of “The MBG Food Poisoning”
27

28. Small Data Sampling for LLM Context
28
Summary
Data Sampling

29. LLM-based Analysis on MBG using SCCT Framework
29
Small data as
context
Prompt
Result
Predefined
Prompts

30. “SCCT” Analysis of MBG Food Poisoning
Short Version
Executive Summary
•17,339 conversations monitored (online news & social media).
•Strong negative sentiment (5,669 negative vs 2,619 positive).
•Dominant emotions: disgust, fear, surprise.
•Amplified by online news (6,435 mentions) & Twitter (9,878 mentions).
•Classified as Preventable Crisis → High attribution of responsibility to government.
•Required response: Rebuild Strategy to restore trust.
Crisis Type (SCCT)
•Preventable Crisis:
•Negligence of SOPs (“rogue kitchens” not following procedures).
•Lack of hygiene certification (SLHS) in most MBG kitchens (Jakarta & West Java).
•Repeated mass poisonings across provinces → systemic failure.
Public Attribution
•Crisis seen as systemic mismanagement, not an accident.
•High responsibility attributed to BGN (National Nutrition Agency) and government.
•President Prabowo linked directly as program initiator → accountability at top level.
Response Strategies
•❌ Deny / Diminish → counterproductive, worsens anger.
•✅ Rebuild (Main Strategy):
•Apology: sincere, empathetic, unconditional.
•Corrective Action: close unsafe kitchens, enforce certification, hotline, Presidential decree on MBG governance.
•Compensation: cover victims’ medical costs.
•➕ Bolstering (Supportive):
•Reminder of noble goals (anti-stunting, children’s nutrition).
•Appreciation to doctors, teachers, parents handling crisis.
Key Messages
•Acknowledgment & Empathy: “We take full responsibility and deeply apologize.”
•Transparency: “No food distribution until 100% certified safe.”
•Long-term Commitment: Reform MBG governance with experts, parents, BPOM.
Channels & Stakeholders
•High-level press conferences (BGN, MoH, BPOM).
•Mainstream online media (Kompas, Tempo, CNN Indonesia).
•Official government social media (Kemenkes, BGN).
•Collaboration: doctors’ associations, academics, NGOs, parents’ committees.
30

31. Example of Constructive Narrative on X
31
Tweet 1“The recent food poisoning incidents in
the Free Nutritious Meal (MBG) program
have raised public concern. Let’s take a
look at what the data tells us.”
(Image: sentiment graph → majority
negative with disgust & fear emotions)
Tweet 2“Out of 17,339 conversations across
online and social media, 5,669 were
negative while only 2,619 were positive.
Dominant emotions: disgust, fear, and
surprise. This shows the public demands
clear and convincing answers.”
Tweet 3“Crisis analysis using the SCCT
framework indicates: the public
perceives this as a Preventable Crisis.
Not an accident, but a systemic failure
that could have been avoided.”
Tweet 4“In crisis communica`on theory, the
appropriate response is the Rebuild
Strategy:
! Sincere apology
" Correc`ve ac`ons (kitchen audits,
hygiene cer`fica`on)
# Compensa`on for vic`ms.”
Tweet 5“Initial steps have been taken: BGN has
apologized, problematic kitchens were
closed, certifications are being
accelerated. But the public still needs
consistent proof before trust can be
restored.”
Tweet 6 (closing)“The MBG program’s goal is noble:
reducing stunting and improving child
nutrition. To keep this vision alive,
transparency and real corrective action
must be prioritized. The public needs
reassurance with data, not just
promises.”

32. Example of Drone Emprit Thread on X
32

33. Drone Emprit Notes Generated by AI
33

34. From Ivory Tower to Newsfeed
Academics can use any LLM—
ChatGPT, Grok, Gemini, Claude, etc.—
as a wri?ng partner to turn complex
research into accessible posts for the
public.
Why it helps
•Clarity: Strip jargon, keep precision.
•Speed: DraE mul?ple formats in
minutes (X/LinkedIn threads, blog
posts, short videos).
•Reach: Meet audiences where they
are with relatable language and
examples.
Simple workflow (5 steps)
•Brief the model: give your thesis, audience, goal, and
3–5 key facts/citations.
•Distill: ask for a plain-English summary at a Grade-
8–10 reading level.
•Localize: add local context, analogies, and everyday
examples.
•Format: request platform-ready outputs (tweet
thread, LinkedIn post, 60-sec video script).
•Verify: fact-check, add references, and human-edit
tone before publishing.
Guardrails
•Provide sources to reduce hallucinations; never ask
the model to invent data.
•Maintain disclosure when appropriate (e.g., “draft
assisted by AI”).
•Keep institutional and ethical standards (privacy,
PDP/GDPR, research integrity).
34

35. Call to Action for Academics
35
Don’t stop at journals → be present in public debates with
dataDon’t stop
Research must have dual purpose: scientific + real-time
literacyHave
Become public intellectuals powered by data, like
MIT/Stanford scholars on social mediaBecome

36. Key Messages
36
Data must live in the
public space – not only in
journals or classrooms.
Academics must be
drivers, not spectators –
but engagement should be
safe and constructive.
Safe Strategy: Adopt
principles of constructive
journalism → focus on
data, frame problems with
solutions, and use
language that educates
rather than provokes.

37. Final QuesTon
If not us—the academics and
researchers—then who will lead the
data-driven public literacy movement?
37

38. THANK YOU
Ismail Fahmi, PhD.
Email: [email protected]