The document summarizes the process an IR team took to develop a solution for SOCPAC PsyOps analysts to better identify and track political influence campaigns. The team initially focused on developing a counter-messaging tool but pivoted after learning this was not realistic. They then worked to understand the problem by interviewing experts and mapping the analysts' workflow. Their final minimum viable product helps analysts identify narratives and accounts spreading disinformation earlier in the process to address a key pain point. The team now has support to further develop the narrative tracking tool and is seeking funding.
1. Katie Jonsson
IR 21’
Dylan Junkin
IR 21’
Will Buchanan
CS 21’
Sally Egan
IR + MCS 22’
ElectionWatch
138 Interviews
Sponsor Problem Statement:
US allies and partners need a
tool to improve the resiliency of
their information environment
and to protect the integrity of
their political elections from
outside interference.
Current Problem Statement:
SOCPAC PsyOps analysts need an
open-source tool to improve the
transparency of the information
environment and track the Chinese
Communist Party’s political influence
campaigns.
2. We interviewed 138 people across government,
industry, and academia
71 Interviews 38 Interviews 29 Interviews
3. “Disinformation is
the next wave of
confrontation
between
authoritarian
regimes and the
United States.”
Senior Official,
Cyber Command
4. But, we had a false start and tackled the
solution before the problem
VALUE PROP
Identify,
attribute,
map, and
respond
BENEFICIARIES
● SOCPAC
● Cyber
Command
● Social media
platforms
● Southeast
Asian
governments
● Focused on solution rather than
problem
● Trouble finding analysts. Too
many ‘high level’ interviews
“We need to identify
malign influence in
elections”
“We need a tool to
directly combat false
information coming from
China and Russia”
-Senior U.S. Government Officials
5. So we restarted and tried to understand the
problem
Week 1 Week 5 Week 9+
Too focused on solution
instead of problem
Morale
New Strategy: Find the subject matter experts
while we schedule to meet with our
beneficiaries!
6. Learning how and why adversaries spread information
Adversaries promote
their own image and
undermine democracy,
using inauthentic and
authentic tactics.
“Policy decisions are made at the
highest levels. This allows
[adversaries] to give broad goals
but no specific implementation
requirements.”
- Glenn Tiffert, Hoover Institution
15
Subject Matter
Experts
Media
Messaging
Academia
Social
Media
Politics
Business
Rest of
Society
Rest of
Society
Adversary
Government
7. Week 1 Week 5 Week 9+
Too focused on solution
instead of problem
Pivot!
We pivoted after learning that influence campaigns target
non-election events and contain truthful elements
We understood the
problem, but not
who owns it
What’s Next?
Election disinformation → political influence
8. Identifying our Beneficiary
Insights:
Overlapping network of
responsibility
Differing approaches
SOCPAC is the Center
of the Spiderweb
Special Operations Command Pacific
Psychological Operations Analysts
State
Department
National
Security Council
Intelligence
Agencies
CIA, FBI, NSA
J39: Information Operations
Psychological Operations
Analysts
J3: Operations
SOCPA
C
INDOPACOM
Department of Defense
CYBERCOM
9. We mapped out the workflow of J39 analysts
monitoring and countering influence
Attribute
“Where did it come from?”
Pain Point: Data sharing
between classification
levels
Goal: Finding the nation/group
behind the narrative
Identify
“Have we been attacked?
Goal: Spotting narratives
targeting US and allies before
they gain traction
Pain Point: Identifying
narratives on social media
Respond
“What should we do about it?”
Pain Point: Counter
messaging is too slow
Goal: Publishing truthful
messages via embassies to
counter the threat
11. KEY PARTNERS
● Policy and Security
specialists at
Facebook, Twitter,
Google, BlackBird,
Primer,ai
● SOCPAC
● Cyber Command
● Southeast Asian
governments
● Cyber policy
research initiatives
(DFRL, SIO)
● Civil Society
● Ambassadors and
PAOs at US
Embassies
● Global Engagement
Center
● State Department
7th Floor (S, P, or R)
KEY RESOURCES
● Twitter APIs
● Election security experts with
understanding of targeting tactics
● Mass link/photo/post
scanning/tracking algorithm
● Student engineers
KEY ACTIVITIES
● Identify and map Chinese to increase
political transparency and allow faster
ingestion of large quantities of data data
MISSION ACHIEVEMENT/IMPACT FACTORS
● Improved transparency on social media platforms regarding disinformation and
influence campaigns
● Ensure allies in Southeast Asia are aware of current Chinese influence
DEPLOYMENT
BUY-IN &
SUPPORT
MISSION BUDGET/COST
● Fixed: software design and engineering,
● Variable: content moderators, cooperation with technology/social
media companies
Once Mission-Fit was achieved, we shifted to deployment
VALUE
PROPOSITIONS
Create pre-
approved
messages so
analysts can
respond quicker
BENEFICIARIES
J-39
Psychological
Operations
Analysts
12. Potential deployment plan in Southeast Asia
Testing our tool where the adversary is testing their tactics
Demo with Special
Operations analysts
Initial testing with US
Embassies in region
Engage with foreign
governments and
society
Demo with Special
Operations analysts
13. 13
After pursuing counter messaging for weeks,
we learned it needs a policy solution, not a
tech solution
“Not sure anyone
can solve
bureaucratic
problems. Getting
messaging out is
strategic, but this
seems like an
impossible task”
US Embassies
Policy Directives
Messaging
14. We can add value through
counter-messaging!Morale
Week 1 Week 5 Week 9+
Counter messaging was
unrealistic. We asked
ourselves: Where can we
add value?
Went back to analyst
workflow to find pain
points to address
Too focused on solution
instead of problem
Pivot!
15. Pivoted from sending counter messages
to identifying narratives
15
Respond Identify
“Our biggest challenge is the
initial assessment of
disinformation… Everything
we have requires having
keywords in order to be
searching for information.”
-SOCPAC Analyst
16. Our final MVP helps analysts identify narratives earlier
“The concept we are talking
about is unique... at the moment
we have nothing that mirrors its
functions…”
- SOCPAC Analyst
17. Identify narratives and
accounts before they
generate significant
attention
What’s Next?
138
Here is where we are now
We can add value through
counter-messaging!
Too focused on solution
instead of problem
Pivot!
18. We are building our narrative tracking
tool and we have support
● Developing partnerships
● Seeking initial grant
funding
● Addressing data access
and competitors
19. We are applying for H4X Funding
The team is pursuing H4X
funding and will be hiring
software engineers this
summer
A big thank you to Major Morton, Lt. Col Liebreich, MG Hager, MG
Braga, Capt. Brunson, Lt. Col Waters, Reece Smyth, Robert Locke, Nini
Moorhead, Daniel Bardenstein, Valeria Rincon, and the teaching team.
The team can be contacted at:
djunkin@stanford.edu
kjonsson@stanford.edu
wbuchan@stanford.edu
egansj18@stanford.edu
21. Reduces time to identify narratives
Runs in background 24/7
Dynamic ability to add/delete actors
Unclassified and open-source
Coordination across analysts
Our MVP addresses key challenges and
integrates with existing workflow: