Elections depend on a reliable information environment where voters can access accurate details about candidates, voting rules, and results. While coordinated efforts to distort this environment are not new, modern digital tools have radically changed their speed, scale, and targeting precision. Today, these operations go beyond simple lies; they involve sophisticated information manipulation designed to exploit specific community vulnerabilities and undermine institutional trust.
This resource explains how these patterns unfold, examines how they have manifested globally, and outlines actionable steps for different actors.
How it could unfold
False claims about registration requirements or eligibility circulate, sometimes through direct messaging at scale. AI-generated ‘deepfake’ audio or video of a candidate or official surfaces to smear them or sow doubt. Spoofed election-authority accounts, websites, or ‘official’ notices push fake rules or deadlines.
False information about polling locations, hours, or ID rules spreads quickly, confusing voters at the moment they try to cast a ballot. Viral, often AI-altered images or clips claim machine failures, ballot ‘dumping,’ or fraud at specific sites. Fake ‘official’ alerts tell targeted communities their polling place has closed or moved.
Unverified claims about counting or machine failures spread before official results are confirmed, creating pressure to contest outcomes. Deepfakes or doctored ‘evidence’ of fraud circulate to justify rejecting the result. Ordinary process steps (curing ballots, moving equipment) are selectively edited and reframed as proof of wrongdoing.
Authentic evidence of real problems can be dismissed as fabricated, making accountability harder.
Perpetrators and targets
Who drives the behavior
- State-Sponsored Actors (Foreign or Domestic): State intelligence services or state-linked entities — a foreign government targeting an adversary, or a government targeting its own public — aiming to weaken rivals, polarize public debate, or reduce trust in democratic institutions.
- Domestic Political Operations: Campaign staff, hyper-partisan political action committees, or hired public relations firms seeking a domestic electoral advantage by depressing opponent turnout or muddying a candidate’s record.
- Commercially Motivated Networks: Content farms and clickbait operators who amplify highly emotional, divisive, or false narratives purely to drive platform engagement and generate advertising revenue.
- Ideological & Fringe Communities: Loose networks of online activists, conspiracy theorists, and automated bot accounts who organically or semi-coordinately amplify false claims to align with their worldview.
Who it is directed at
- Marginalized & Immigrant Communities: Often targeted with tailored, foreign-language disinformation because they may rely on closed messaging apps (like WhatsApp or Viber) and may have less access to official, mainstream election communications.
- Election Administrators & Poll Workers: Targeted with coordinated harassment and doxxing to create institutional stress, force staff resignations, and disrupt the physical operational capacity of election offices.
- Undecided or Low-Information Voters: Targeted with hyper-specific micro-targeting (often regarding changing polling locations or ID rules) to confuse them at the moment of voting or discourage them from participating entirely.
- The Mainstream Media Grid: Targeted with ‘tainted leaks’ or deepfakes specifically timed to exploit the fast-paced breaking news cycle, forcing journalists to inadvertently amplify false narratives simply by reporting on their existence.
Examples from around the world
Leaked documents released during a legal media blackout
Hours before France’s legal pre-election media silence period, a large document trove — some genuine, some fabricated — was released online purporting to show misconduct by a presidential candidate. French media largely declined to amplify unverified material during the blackout period. International platforms were not bound by French law and distributed it widely.
Lesson: Pre-existing legal frameworks for media conduct can reduce the window for false claims to take hold, but they have no reach beyond national borders.
Sustained claims of machine fraud leading to post-election unrest
Brazil’s tightly contested 2022 presidential election was accompanied by a sustained campaign, primarily through WhatsApp, claiming that electronic voting machines were compromised. The Superior Electoral Court took an unusually active role fact-checking claims in real time. After results were certified, supporters of the losing candidate blockaded highways and military bases for weeks, partly citing false information absorbed across the campaign period.
Lesson: Official fact-checking may not reach people embedded in closed messaging networks. Post-election civic disruption can follow sustained pre-election false claim campaigns even when the underlying electoral process is sound.
AI-generated audio released inside a legal response blackout
Two days before Slovakia’s parliamentary election and within the legal pre-election silence period, an audio clip circulated on social media appearing to show a candidate discussing vote-rigging. Independent analysis suggested the audio was synthetically generated. Because the silence period prevented formal responses, the clip spread without counter-messaging for the final days of the campaign.
Lesson: Synthetic media timed to fall inside legal response windows can be particularly effective. Pre-election authentication infrastructure matters more as AI-generation tools become widely accessible.
One-on-one automated messaging replacing broadcast disinformation
The 2022 Philippine presidential election saw a shift toward personalized automated messaging through social messaging networks such as Viber and WhatsApp, with bots engaging individual voters in direct conversations and sharing tailored false information about candidates and voting procedures. This approach is harder to detect than public viral posts because it leaves no visible public footprint for watchdog groups to audit.
Lesson: Monitoring public platforms for false claims misses coordinated activity happening in private channels. Community-level monitoring by trusted organizations is a more effective early warning system.
Slow results reporting exploited as evidence of fraud
Kenya’s 2022 presidential election involved a close result and a prolonged results transmission process. In the absence of timely official communication, false narratives about hidden results and system manipulation spread rapidly on social media, including photoshopped tally sheets. The electoral commission had not successfully pre-positioned public communications explaining why results take time to transmit and verify.
Lesson: Delayed results are normal and often a sign of careful counting, but without pre-positioned public explanation, the gap fills with false claims. Communicating expected timelines in advance is one of the highest-leverage preventive actions available to election administrators.
Signals to watch for
| Before | Identical or near-identical messages about voting rules appearing across platforms with no clear organic origin |
| Before | A public figure preemptively claims upcoming videos or recordings will be fabricated before any such content exists |
| Before | Unusual spike in election office calls about ID or document requirements that don’t match official guidance |
| Before | Foreign-language content circulating in some communities with no traceable domestic origin |
| Before | Coordinated harassment of named election officials on social media escalating in volume |
| Before | Multiple platforms simultaneously reducing election-specific content moderation |
| During | Viral claims about polling site closures or machine failures not reflected on any official channel |
| During | Reports from community organizations of voters receiving direct messages with false location or timing information |
| After | Results reporting slows and unverified explanation narratives fill the gap before officials communicate |
| After | Authentic evidence of problems publicly dismissed as fabricated without technical analysis to support that claim |
| Ongoing | Candidates or parties pre-announcing they will only accept results if they win |
| Ongoing | Spike in AI-generated political content volume around key dates or candidate events |
How risk can escalate
Routine partisan disputes or isolated false claims, correctable through standard fact-checking. Usually limited in impact.
Coordinated false narratives about voting logistics or eligibility spread in specific communities. May suppress turnout or cause confusion but is more localized.
Synthetic media (deepfake) goes viral about fraud, misconduct or other issue; verified evidence successfully dismissed as fabricated. Could erode public confidence and cause a certain level of election disruption.
False claims trigger physical disruptions or widespread refusal to accept certified results. Could lead to constitutional crisis, closing of polling sites due to security threats, officials refusing to certify, widespread panic, election violence.
What different actors can do
Lawmakers & policymakers
- Require transparency for political advertising and platform conduct. Mandate disclosure of who pays for political advertising, labeling of AI-generated political content, and regular platform reporting on coordinated inauthentic behavior — with the resources and penalties needed to make the rules enforceable rather than voluntary.
- Set narrowly drawn rules for deceptive synthetic media in elections. Require disclosure or labeling of materially deceptive AI-generated depictions of candidates and officials, drafted narrowly to protect satire, reporting, and legitimate political speech.
- Fund a permanent, neutral election-communications capacity so authoritative information infrastructure exists year-round rather than being improvised each cycle.
- Fund sustained civic and voter education so people understand how elections are actually run — how votes are cast, counted, and certified, and why results take time — before false claims fill the gap.
- Protect and resource independent fact-checking and media-literacy programs without controlling their content, so resilience is built in civil society rather than dictated by the state.
Election administrators
- Authenticate official communications. Use verifiable watermarking or digital signatures on official video and audio so voters can confirm what they are seeing is genuine.
- Build a single authoritative information hub — multilingual if needed and prominently promoted — that voters can be directed to before false claims emerge, not after.
- Explain expected timelines in advance. Pre-position public communications about why results take time to report. This neutralizes a common false claim before it starts.
- Maintain direct contacts with platform trust and safety teams for rapid escalation of coordinated false claims.
- Issue factual updates at regular intervals, not only in response to specific false claims. Silence creates space for false narratives to fill.
- Respond to viral false claims quickly and specifically. Generic denials are less effective than addressing the specific claim with specific evidence.
- Establish a verification window protocol — a publicly agreed standard for how long before a claim about official conduct is labeled unverified vs. confirmed or denied.
- Invest in permanent election communications infrastructure — maintained year-round so credibility is established before it is needed.
- Build standing partnerships with civil society and trusted messengers so verified information and early warnings flow both ways and reach communities through channels they already trust.
- Train staff in narrative management so rapid accurate responses can be issued without waiting for lengthy approval chains.
Civil society organizations
- Pre-bunk known false narratives before they spread. Identify patterns from prior cycles and deploy counter-messaging before claims emerge.
- Establish direct lines to election administrators so accurate information and emerging false-claim threats can be shared quickly in both directions, before and during the vote.
- Build trusted messenger networks with local media, faith leaders, and ethnic media outlets who can push accurate information quickly to targeted communities.
- Monitor closed messaging platforms, not just public social media. Coordinated false claims increasingly spread through private channels first.
- Document patterns systematically. Evidence of coordinated inauthentic behavior requires documentation across many instances — single incidents are rarely actionable alone.
- Activate rapid response networks within hours of a false claim going viral. Trusted messengers with established credibility counter narratives faster than official channels in many cases.
- Escalate to platform trust and safety teams with documented evidence of coordinated behavior, particularly on Election Day.
- Support affected communities directly with accurate voting information through trusted local intermediaries, especially where false claims have already spread.
- Invest in media literacy programs year-round, not only in election periods. Sustained civic investment in information literacy provides resilience that reactive fact-checking cannot replicate.
- Build coalitions of trusted messengers across communities well before election cycles — credibility cannot be established in the moment it is needed.
Community members
- Verify voting information at the source. Use your official state or local election authority website, and not social media, for polling locations, hours, and ID requirements.
- Share accurate information proactively with friends and family before false claims take hold. Personal networks are more trusted than official channels for many people.
- Pause before sharing anything about voting rules or results. False claims about elections spread fastest in the first hour — waiting for confirmation from an official source costs little and reduces harm significantly.
- Report, don’t amplify. If you see coordinated false claims about voting, report them to the platform rather than sharing with commentary because even critical commentary increases reach.
- If you encounter problems voting, report them. Election protection hotlines and civil society observers can only act on problems they know about.
- Support local journalism. Local reporters who cover election administration year-round are one of the most effective checks on false claims about local elections.
Voters
- Learn how your local election actually works — how ballots are cast, counted, and certified in your area, and how long results take — so you can recognize false claims when you see them.
- Look up your polling location, hours, and ID requirements on your official state election website at least one week before the election — not on the day itself.
- Confirm last-minute messages about changed polling locations or new requirements on official websites.
- Know that slow results are normal. Results take time to count accurately. Delays in reporting are not evidence of manipulation.
Examples of effective response
A prepared civic ecosystem absorbed a surge of AI-generated content
Ahead of and during Taiwan’s January 2024 election, high volumes of AI-generated audio, images, and text — much of it attributed to foreign-linked influence operations — circulated online. Rather than relying on takedowns alone, Taiwan drew on years of prior investment: crowdsourced fact-checking services and independent fact-check organizations, government rapid-rebuttal teams that answered false claims within hours (often with plain or even humorous corrections), and media literacy embedded in the school curriculum. The false content spread but did not demonstrably shift results or public behavior.
Takeaway: Standing capacity beats reactive cleanup. The ‘invest in media literacy year-round’ and ‘activate rapid-response networks’ recommendations work best when built long before the election rather than stood up in the moment.
Enforceable platform obligations and a shared rapid-alert network for an election year
For the 2024 European Parliament elections, the EU used its platform-accountability law to require the largest platforms to assess and mitigate election-related information risks, published specific election guidelines (including expectations on labeling AI-generated content and surfacing official information), and coordinated a rapid-alert network among member states to share emerging manipulation campaigns. Because the obligations carried the possibility of real penalties, the platform commitments had weight rather than remaining voluntary. In practice, though, enforcement was uneven across member states, and platforms still struggled to label AI-generated content in smaller regional languages — regulatory leverage, not a silver bullet.
Takeaway: The ‘enforceable transparency and platform-accountability rules’ recommendation in practice: obligations with penalties behind them give official channels leverage that voluntary cooperation alone does not — but only where the rules are actually enforced, evenly, across jurisdictions and languages.
The election authority fact-checked in real time and partnered with platforms
Facing a sustained messaging-app campaign claiming its electronic voting machines were rigged, Brazil’s Superior Electoral Court ran an unusually proactive counter-disinformation program in 2022 — an official fact-checking hub, fast public rebuttals, and agreements with messaging and social platforms to limit mass-forwarding and surface authoritative election information. The measures did not prevent post-election unrest, but they contested false claims in the open throughout the campaign rather than ceding the ground.
Takeaway: The ‘single authoritative information hub’ and ‘rapid, specific rebuttal’ recommendations run directly by the election authority — necessary, though a reminder that real-time correction alone cannot undo narratives absorbed over months.
A national media-literacy curriculum built long-term resistance
Finland has for years embedded critical-thinking and media-literacy skills across its schools and public education — teaching students and adults to identify manipulation, check sources, and understand how false narratives are engineered. Independent European assessments have repeatedly ranked Finland’s population among the most resilient to misinformation, an advantage that predates any single election and does not depend on reacting to specific claims.
Takeaway: The clearest form of the ‘invest in media literacy year-round’ recommendation: resilience built into education pays off across every election cycle, not just one.
An official ‘rumor control’ hub and a joint security statement pre-empted fraud claims
During the 2020 election, the federal cybersecurity agency ran a public ‘rumor control’ resource that stated common false claims about voting and results and answered them directly, and a coalition of federal and state election-security officials issued a joint statement describing the election as secure and finding no evidence that voting systems were compromised. Both aimed to put authoritative, specific information in front of voters quickly rather than letting false narratives about the count stand unanswered.
Takeaway: The ‘single authoritative information hub’ and ‘rapid, specific correction’ recommendations in official form — though it also shows such rebuttals work best when established and trusted before results are contested, not improvised midstream.