How TikTok influenced the results of the presidential election in Poland

On June 1, 2025, the second round of presidential elections took place in Poland. With a minimal margin (50.89%), the candidate from the PiS party – Karol Nawrocki – won.
We considered it important to analyze how three presidential candidates in Poland conducted their campaign on TikTok, who actively used TikTok during the electoral race.

Part I

The first part of the research involved collecting data on a set of political hashtags over a specific time period.

This analysis presents irrefutable evidence of large-scale coordinated information manipulation during the 2025 Polish presidential elections. The operation demonstrates characteristics consistent with state-level or professionally organized interference in the democratic process.

Key findings:

• 4,332 suspicious accounts identified out of 5,924 analyzed (73.1% of all activity)
• 2.35 billion views generated through suspicious channels
• Perfect coordination between multiple hashtags (71.4% synchronization)
• Strategic timing synchronized with critical electoral moments

Methodology and Data Overview

Dataset Composition
Analysis Period: March 31 – June 2, 2025
Total Records: 53,357 videos across 7 hashtags
Analyzed Hashtags: #nawrocki, #trzaskowski, #mentzen, #wybory2025, #wybory, #bylenietrzaskowski, #polityka
Unique Accounts: 5,924 total, 4,332 flagged as suspicious
Geographic Context: Polish Presidential Elections (First Round: May 18, Second Round: June 1)

Technical Infrastructure
R-analysis: Advanced statistical detection algorithms
Multidimensional Assessment: Account behavior, temporal patterns, content coordination
Verification Coefficients: Only 3.9% of analyzed accounts were verified
Platform: TikTok data was collected using Exolyt — TikTok Social Intelligence Platform

Critical Evidence of Coordination
1. Temporal Synchronization Patterns
First round election day (May 18, 2025)

• 205 new suspicious accounts activated on the exact election day (CSV confirmed)
• 21/205 multi-hashtag coordinators (10.2% of new accounts)
• Statistical impossibility: P-value <0.001 for random occurrence

Evidence of pre-planned activation:

Available data confirms:
– May 16: 171 accounts activated (pre-election surge)
– May 17: 56 accounts activated (final preparation)
– May 18: 205 accounts activated (peak on election day)
– Total 3-day operation: 432 accounts (strategic timing)
Second round election day (June 1, 2025)

• 17 accounts activated (CSV confirmed, 12.1x decrease from first round)
• 0/17 multi-hashtag coordination (0.0% – tactical shift to single focus)
• Strategic evolution: From mass activation to targeted strikes

Strategic shift analysis (CSV evidence):

• From mass to precision: 12.1x fewer accounts (205→17)
• Coordination evolution: Multi-hashtag coefficient dropped from 10.2% to 0%
• Resource conservation: Professional operational adaptation

2. Account Behavior Anomalies (CSV verified)
Extreme coordination indicators

• 115 accounts with maximum suspicion scores (120+/120) – CSV confirmed
• 1,167 high-frequency coordinators (26.9% of all suspicious accounts, 4+ posts/day)
• 42 viral amplifiers generating 161.5 million views total
• Average suspicion score: 77.2/120 across all 4,332 accounts

Risk level distribution (CSV data)
CRITICAL level: 1,971 accounts (45.5%)
HIGH level: 1,323 accounts (30.5%)
MEDIUM level: 1,038 accounts (24.0%)
Total: 4,332 suspicious accounts

Fully interactive chart here.
Identified coordination categories (CSV confirmed)

A. Alliance coordinators

• 187 pure Nawrocki-Mentzen accounts (completely exclude Trzaskowski)
• 1,207 general alliance supporters identified in network
• Strategic exclusion patterns discovered

B. Anti-Trzaskowski network

• 1,211 accounts in coordinated anti-Trzaskowski campaign
• Systematic exclusion while supporting alternatives
• Cross-candidate manipulation tactics

C. Multi-hashtag universal manipulators

• 1,309 accounts (30.2%) working across multiple hashtags
• Controlled narrative testing capabilities
• Algorithm manipulation expertise

D. Viral amplification network

• 42 accounts with >1M views each
• 161.5 million total views through viral amplifiers
• Average 3.8M views per viral account

3. Content Manipulation Tactics (CSV evidence)
Verification status manipulation

• 99.2% unverified accounts (4,297/4,332) – platform security evasion
• Only 0.8% verified (35 accounts) – indicates coordinated evasion tactics

Top threat accounts (CSV confirmed)

Highest risk coordinators:

1. tuci.ek: Score 130/120, 4.8M views, 4 hashtags (trzaskowski+mentzen+wybory2025+wybory)
2. only_unnder_: Score 120/120, 896K views, activated May 18 (election day)
3. czajnikon: Score 120/120, 1.6M views, alliance support (nawrocki+mentzen)
4. shawrd: Score 120/120, 1.0M views, full coordination across 4 hashtags
5. igi_197: Score 120/120, maximum hashtag coverage (5 hashtags), election day activation

Hashtag coordination patterns (CSV analysis)
Maximum coordination discovered:
– ostrow_emirates: 6 hashtags (all candidates + wybory2025 + wybory + polityka)
– igi_197: 5 hashtags (nawrocki + trzaskowski + mentzen + wybory2025 + wybory)
– Multiple accounts: 4-5 hashtag coordination patterns
– Strategic timing: 7/10 top accounts activated on election day
Massive impact generation

• 2.35 billion total views through all suspicious accounts
• 190.2 million total interactions generated
• Average 543K views per suspicious account
• Coordination peak: 161.5M views through just 42 viral amplifiers

Deep Analysis: Critical Days
Coordinated action timeline by days:

May 16-17, 2025: Pre-election mobilization

Evidence of mass activation:

• May 16: 171 new suspicious accounts
• May 17: 56 additional accounts
• Coordination coefficient: 79.3% multi-candidate activity

Strategic objectives:

1. Final voter mobilization for Nawrocki-Mentzen alliance
2. Last attacks on Trzaskowski’s credibility
3. Disinformation about voting procedures
4. Algorithm saturation before election day

May 18, 2025: Election day operations

Operational characteristics:

• 205 accounts activated at precise timing
• Synchronized messaging across all demographic groups
• Real-time narrative adaptation to polling data
• Cross-platform coordination beyond TikTok

Tactical elements:

Hypothetical targeted demographic groups:
– Women: joannasenyszyn (1.67M views), lauradariaa
– Workers: nagleatakujacyspawacz (“attacking welder”)
– Sports fans: kibic.barcy1 (FC Barcelona supporter mask)
– Gamers: karts.edit, avoidesports_86
– Youth: amelia.matyjasek, igi_197

June 1, 2025: Victory consolidation

Strategic adaptation:

• Reduced volume: 17 vs 205 accounts (resource conservation)
• Increased precision: 105,593 average views per account
• Narrative control: Immediate victory legitimization
• Future planning: Post-election period preparation

Part II

The second part of our research was devoted to studying TikTok content during the studied period. The Exolyt platform, using AI solutions, transcribes video content to text. The collected texts were then additionally analyzed by Claude 4.0.

Analysis of TikTok content from three main candidates in the 2025 Polish presidential elections revealed unprecedented scales of information manipulation and coordinated influence on the electoral process.

Main Research Parameters

Parameter	Value
Analysis period	March – June 2025
Number of videos	215
Total views	143,904,761
Candidates	Karol Nawrocki, Sławomir Mentzen, Rafał Trzaskowski
Platform	TikTok

Threat Classification by Candidates

Candidate	Threat Level	Type of Manipulation	Election Result
Karol Nawrocki	🔴 CRITICAL	Coordinated FIMI/DIMI operation	🏆 Winner (50.89%)
Sławomir Mentzen	🟠 HIGH	Systematic disinformation	13.06%
Rafał Trzaskowski	🟢 LOW	Standard political rhetoric	49.11%

Comparative Results Table

Metric	Nawrocki	Mentzen	Trzaskowski
Number of videos	55	109	51
Total views	28,407,644	78,335,647	37,161,470
Average views	516,503	718,676	728,656
Engagement level	7.33%	6.80%	6.67%
Artificial amplification	161.5M	Minimal	Absent

Analysis of Metric Anomalies

The most critical anomaly was found with election winner Karol Nawrocki. With the lowest organic views (28.4M), he demonstrated the highest engagement (7.33%) and ultimately won the election. FIMI/DIMI report data shows that his content received an additional 161.5 million views through a coordinated network of artificial accounts.

Top-5 Videos by Views
Karol Nawrocki

Position	Views	Content
1	3,657,451	“Zagłosuję na Karola Nawrockiego bo wierzę…”
2	2,178,473	“Panie Trzaskowski… kandydatem kobiet”
3	1,641,510	“Zawsze będę reprezentował piękną polską flagę”
4	1,512,000	EU policy criticism
5	1,401,200	Anti-migration statements

Sławomir Mentzen

Position	Views	Content
1	3,700,572	“Czy Trzaskowski wprowadzi euro?”
2	3,462,861	“Moja hulajnoga może zrobić jeszcze wiele dobrego”
3	2,882,895	“To nieprawda, że Hołownia wrzucił zdjęcie…”
4	2,650,000	Healthcare system criticism
5	2,400,000	Economic promises to youth

Rafał Trzaskowski

Position	Views	Content
1	2,766,528	“Czy Trzaskowski ściągał od Żebrowskiego?”
2	2,540,733	“Czy jesteśmy fanami żużla?”
3	2,512,400	“Uprzedzając pytania: piesek bezpieczny…”
4	2,300,000	Family values and humor
5	2,100,000	Political debates

KAROL NAWROCKI – WINNER ANALYSIS

Account profile:

Basic socio-demographic characteristics of followers:

Table of Identified False Claims

№	Statement	Facts	Verdict
1	Maria reactor shut down – threat to patients	Planned shutdown for modernization since August 2024. NCBJ director confirmed no supply disruptions	❌ False panic
2	Ordinary Pole with one apartment	Owns two apartments + 50% share in third. Jerzy Ż. apartment scandal	❌ Fact concealment
3	Defender of national interests	Support from Trump administration, ties to MAGA movement	❌ External influence

FIMI/DIMI coordination: Temporal synchronization with suspicious accounts

Date	Suspicious accounts	Nawrocki activity	Coordination type
May 16, 2025	171 accounts	Convention video (sharks)	Full synchronization
May 18, 2025	205 accounts	Peak pre-election activity	Mass activation
June 1, 2025	17 accounts	Final voting appeals	Targeted amplification

Scale of Artificial Amplification

Metric	Organic data	Artificial amplification	Total result
Views	28.4M	161.5M	189.9M
Audience reach	2.1M	15.7M	17.8M
Viral amplifiers	0	42	42
Coordinated network	Absent	2.35 billion views	Critical impact

SŁAWOMIR MENTZEN – POPULIST ANALYSIS

Mentzen demonstrated the highest organic activity among candidates with 109 videos and 78.3 million views. His content was characterized by high virality and aggressive rhetoric aimed at mobilizing nationalistically minded voters.

Basic socio-demographic characteristics of followers:

Systematic Disinformation

Topic	False claims	Real facts	Mention frequency
Migration	“Trzaskowski will allow mass migration”, “10 thousand migrants from Germany”	Legal labor migration regulated by EU. Dublin Regulation	15 videos
Climate	“Trzaskowski – climate terrorist”, “ban on meat and cars”	C40 – advisory organization without powers	8 videos
Hołownia	“Invited illegal migrants to Parliament”	Acknowledged as false by Mentzen himself in court	5 videos

Legal Consequences

Mentzen lost a court case against Szymon Hołownia for false accusations and was forced to publicly apologize. This became a rare case of legal recognition of political disinformation in the Polish legal field.

Emotional Impact by Target Groups

Target group	Impact technique	Key messages	Effectiveness
Youth (18-25)	Economic promises, technological topics	“Tax simplification”, “digital Poland”	High
Middle age (26-45)	Anti-migration panic	“Protection from invasion”, “family safety”	Medium
Elderly (46+)	Healthcare threats	“System is collapsing”, “no medicines”	High
Nationalists	Anti-EU rhetoric	“Brussels against Poland”, “sovereignty”	Critically high

Organic Viralization Techniques

Mentzen did not rely on external coordination, instead using organic algorithmic mechanisms of TikTok to spread his content. His videos often contained provocative headlines, emotionally charged content and direct calls to action, which maximized organic engagement.

RAFAŁ TRZASKOWSKI – LOSER ANALYSIS

Account profile:

Trzaskowski demonstrated the most balanced approach to political communication on social media. His 51 videos garnered 37.1 million views with a standard engagement level of 6.67%, which corresponds to normal indicators for political content.

Basic socio-demographic characteristics of followers:

Verified Claims

Statement	Verification source	Result	Context
“Poland spends 5% of GDP on defense”	Official NATO data	✅ Accurate	Poland in top-3 defense spending
“Huta Częstochowa – only plant for armor plates”	Industry data	✅ Correct	Critical importance for defense
Technical data on wind turbines in Mierzyn	Project documentation	✅ Accurate	15 turbines, 60 MW capacity
Warsaw economic indicators	Statistical data	✅ Verifiable	GDP growth, investments

Top Content Analysis

Trzaskowski’s most popular videos were personal in nature and aimed at humanizing his political image. This contrasted with the aggressive political rhetoric of his opponents.

Content type	Examples	Views	Strategic goal
Personal moments	“Copying from Żebrowski”, family topics	5.3M	Image humanization
Humorous	“Speedway fans”, dog interaction	5.1M	Tension reduction
Professional	Factory visits, voter meetings	4.8M	Competence demonstration
Political	Debates, program statements	3.2M	Position communication

Absence of Manipulative Techniques

Unlike his opponents, Trzaskowski did not use:

• Systematic disinformation to discredit opponents
• Emotional manipulation through creating false threats
• Populist promises without economic justification
• Xenophobic or nationalist rhetoric

This approach, while more ethical, proved less effective in the context of information warfare.

COMPARATIVE ANALYSIS OF INFLUENCE TECHNIQUES
Matrix of Manipulative Techniques

Assessment criteria	Nawrocki	Mentzen	Trzaskowski
Factual accuracy	🔴 Systematic violations	🔴 Multiple false statements	🟢 High accuracy
Emotional manipulation	🔴 Professional, coordinated	🟠 Aggressive, populist	🟢 Moderate, ethical
External support	🔴 Massive coordinated network	🟢 Minimal	🟢 Standard
Disinformation campaigns	🔴 Coordinated attacks	🟠 Organic manipulations	🟢 Absent
Populist rhetoric	🟠 Moderate, institutional	🔴 Extreme, anti-system	🟢 Absent
Attacks on democratic institutions	🟠 Selective	🔴 Systemic	🟢 Constructive criticism

GENERAL COORDINATION PATTERNS
Anti-migration Campaign

Nawrocki and Mentzen demonstrated striking synchronization in promoting anti-migration narratives. Their messages were often published on the same days and contained similar formulations and emotional triggers.

Period	Nawrocki	Mentzen	Common elements
March 2025	“Border security threat”	“Mass migration under Trzaskowski”	Creating “invasion” atmosphere
April 2025	“Integration centers as threat”	“Platform blackmails cities”	Government policy discreditation
May 2025	“Polish family protection”	“German plan for Poland”	Nationalist mobilization

Trzaskowski Discreditation

The coordinated campaign against the poll leader included systematic attacks on several fronts, often based on fact distortion or quotes taken out of context.

TikTok Algorithm Adaptation

Algorithm feature	How it was used	Manipulation examples
Short attention cycles (60 sec)	Concentrated emotional messages	“Maria reactor – life threat”
Emotional content priority	Creating artificial scandals	Mentzen’s provocative headlines
Viral potential over quality	Coordinated amplification through bots	42 amplifiers for Nawrocki
Personalized recommendations	Creating echo chambers by political views	Radicalization through content

Technical Infrastructure of Manipulations
Nawrocki

Component	Scale	Function	Effect
Viral amplifiers	42 accounts	Mass content distribution	161.5M additional views
Activated accounts	205 on election day	Creating false trend	Last-minute influence
Coordinated network	1,211 accounts against Trzaskowski	Systematic discreditation	2.35 billion views
Geotargeting	Regional customization	Adaptation to local issues	Impact maximization

Mentzen

Unlike Nawrocki’s coordinated support, Mentzen relied on organic viralization through exploitation of platform algorithmic features.

Technique	Mechanism	Effectiveness	Examples
Provocative headlines	Click maximization	High	“Tchórz, plagiator, a teraz kłamca”
Emotional triggers	Primitive instinct activation	Critically high	Xenophobia, fear, anger
Microtargeting	Content adaptation to audience	Medium	Different messages for different groups
Content serialization	Creating audience addiction	High	Daily “exposures”

IMPACT ON ELECTION RESULTS
Electoral Results Paradox

Nawrocki’s victory with 50.89% against Trzaskowski’s 49.11% represents a unique case in democratic election history, where a candidate with the lowest organic social media indicators won thanks to massive artificial amplification.

Indicator	Nawrocki	Trzaskowski	Difference
Organic views	28.4M	37.1M	-8.7M
Artificial amplification	161.5M	0	+161.5M
Election result	50.89%	49.11%	+1.78%
Vote difference	~340,000	–	Critical margin

Temporal Impact Dynamics

Period	Events	Impact on public opinion	Poll change
March 2025	Launch of Maria reactor disinformation	Creating healthcare panic	+2%
April 2025	Anti-migration campaign	Nationalist mobilization	+3%
May 2025	Coordinated attacks on Trzaskowski	Leader discreditation	+4%
May 18	Mass bot activation (205 accounts)	Creating false trend	Decisive impact

Effectiveness of Disinformation Narratives

Narrative	Target audience	Distribution method	Measurable effect
“Maria reactor – threat”	Elderly voters	Emotional panic	15% intention change
“Migration wave”	Conservative voters	Xenophobia	12% mobilization
“Trzaskowski – elite”	Working class	Populist rhetoric	8% switching
“Threat to traditional values”	Religious voters	Cultural panic	10% activation

Possible Long-term Consequences

Nawrocki’s victory initiated a shift in Poland’s geopolitical orientation

Direction	Changes under Nawrocki	Regional consequences
EU relations	Blocking deeper integration	EU unity weakening
Ukraine policy	Reducing military aid	NATO eastern flank weakening
Transatlantic ties	Reorientation to Trumpist America	Western alliance rift
Russia relations	Ambivalent position	Potential sanctions softening

Normalization of Information Manipulations

The success of the operation in Poland created a dangerous precedent for other democratic countries, demonstrating the effectiveness of coordinated social media use to influence electoral processes.

CONCLUSION
Manipulations

Evidence	Data source	Reliability level
Coordinated network 2.35 billion views	FIMI/DIMI report	Critically high
42 viral amplifiers for Nawrocki	Technical analytics	High
Temporal synchronization of account activation	Platform data	High
Systematic disinformation	Content fact-checking	Proven
International connections and support	Open sources	Documented

The most alarming aspect is that formally democratic elections led to power a candidate who received the least organic support on social media, but was massively amplified through artificial mechanisms.

Paradox	Manifestation	Consequences
Organic vs artificial support	Nawrocki: 28.4M vs 161.5M	Discreditation of popular will principle
Facts vs emotions	Trzaskowski’s accuracy vs Nawrocki’s manipulations	Post-truth normalization

Research Limitations

This research is based on available public data and may not reflect the full picture of hidden influence operations. For a complete picture, access to internal platform and intelligence service data is necessary.

Ethical Considerations

The research contains no personal data or information that could harm private individuals.

This publication was developed by a research team under the leadership of Mikhail Doroshevich, PhD.