Foreign interference: outside governments trying to shape another country’s vote

In plain words

Foreign interference in elections means efforts by foreign governments, or people working for them, to affect another country’s election, either by trying to change voters’ views or by attacking the machinery of voting itself. U.S. intelligence agencies have documented influence campaigns by Russia and Iran in the 2020 U.S. elections, while reporting no indication that any foreign actor altered the technical voting process.

The simple version: a foreign government might pay people to post misleading messages online, leak stolen emails, or try to break into election computers. The first kind tries to change minds. The second tries to change counts. They need different defenses. And there is a trap on both sides: ignoring real attacks is dangerous, but calling every unpopular opinion “foreign” can be used to silence citizens.

The problem, stated precisely

The problem has two parts that should not be blurred:

1. Influence. Overt or covert efforts by foreign governments or their agents to affect candidates, parties, voters, their preferences, or political processes.
2. Interference. The narrower subset aimed at the technical aspects of an election: voter registration, casting and counting ballots, and reporting results.

These definitions come from the U.S. Intelligence Community’s March 2021 assessment.⁠¹ and are used here because they separate attacks on minds from attacks on machinery. The page does not cover domestic misinformation, which is on misinformation, or domestic campaign money, which is on money in politics. Technical election security is covered in depth on security.

Defined terms

• Election influence. As defined by the 2021 Intelligence Community Assessment, “overt and covert efforts by foreign governments or actors acting as agents of, or on behalf of, foreign governments intended to affect directly or indirectly a US election”⁠¹
• Election interference. In the same assessment, “a subset of election influence activities targeted at the technical aspects of the election, including voter registration, casting and counting ballots, or reporting results”.⁠¹
• Hack and leak. Stealing private information and releasing it to shape coverage.
• Foreign national. Under U.S. campaign finance law, foreign principals and certain non-citizens who are not lawful permanent residents.⁠²
• Disclosure regime. A legal approach that requires people acting for foreign governments to register and report, rather than banning the activity outright.⁠³

Evidence

The 2020 U.S. assessment

The Intelligence Community Assessment dated March 10, 2021, and declassified March 15, 2021, reached five key judgments:⁠¹

• “We have no indications that any foreign actor attempted to alter any technical aspect of the voting process in the 2020 US elections, including voter registration, casting ballots, vote tabulation, or reporting results.” It added that it would be difficult to manipulate election processes at scale without detection, including through post-election audits, and that Iran and Russia “spread false or inflated claims about alleged compromises of voting systems to undermine public confidence.”
• Russian President Putin authorized, and Russian government organizations conducted, influence operations aimed at denigrating one candidate, supporting another, undermining confidence in the electoral process, and exacerbating divisions. The assessment noted: “Unlike in 2016, we did not see persistent Russian cyber efforts to gain access to election infrastructure.”
• Iran carried out a covert influence campaign intended to undercut the reelection of the incumbent, undermine confidence, and sow division, authorized by Supreme Leader Khamenei.
• China “did not deploy interference efforts and considered but did not deploy influence efforts intended to change the outcome,” though the National Intelligence Officer for Cyber assessed China took some steps against the incumbent’s reelection.
• Other actors, including Lebanese Hizballah, Cuba, and Venezuela, took smaller steps.

Two points matter for readers from any side. Foreign influence efforts in 2020 aimed at both major candidates, depending on the foreign actor. And the assessment found influence, not altered counts.

Measured effect of an online influence campaign

The question of effect is where evidence is most needed and hardest to get. Eady, Paskhalis, Zilinsky, Bonneau, Nagler, and Tucker, in Nature Communications (January 2023), linked a three-wave survey of 1,496 U.S. respondents to their Twitter feeds to study exposure to posts from Russia’s Internet Research Agency during the 2016 campaign.⁠⁴ They found exposure was highly concentrated: “Only 1% of users accounted for 70% of exposures.” Strong Republicans saw about nine times as many Russian posts as Democrats or independents. Respondents saw about 25 times more posts from national news media and 9 times more from politicians than from Russian accounts in the final month. The authors found no meaningful relationship between exposure and changes in issue attitudes, perceived polarization, or vote choice, while cautioning that their data are observational.⁠⁴ They also estimated about 32 million Americans were potentially exposed.⁠⁴

This study covers one platform and one campaign. It does not measure the effect of hack-and-leak operations, of attacks on infrastructure, or of foreign efforts to erode trust after an election.

Elsewhere: an election annulled

In Romania, the Constitutional Court annulled the first round of the 2024 presidential election on December 6, 2024, two days before the scheduled runoff, after intelligence documents were declassified stating that Russia had run a coordinated online campaign to promote one candidate, according to a secondary summary of the events.⁠⁵ The court had confirmed the first-round results on December 2 and reversed itself on December 6.⁠⁵ The case is a test of how far a democracy may go in response: supporters saw protection of the vote, while critics worried about courts voiding elections on intelligence findings voters could not fully see.

Causes and competing explanations

• Low cost, high reach. Online platforms let a foreign actor reach large audiences cheaply, though Eady and colleagues found reach concentrated among a small share of users.⁠⁴
• Domestic division as fuel. Influence campaigns aim to “exacerbate sociopolitical divisions”.⁠¹ They work with material a society already supplies, which links this problem to polarization.
• Trust as the target. The 2021 assessment describes false claims about compromised voting systems intended to undermine confidence.⁠¹ A foreign actor does not need to change a vote if it can persuade people the count is untrustworthy; see trust in elections.
• Strategic motives. The assessment says China weighed the risk of being caught against expected gains and chose not to deploy influence efforts to change the outcome.⁠¹ Foreign behavior responds to costs.

Who is harmed

Voters are harmed when they are deceived about who is speaking to them. Candidates and parties are harmed when stolen material is used against them. Election officials are harmed when false claims of compromise target their work.⁠¹ And everyone is harmed if the public loses the ability to tell foreign manipulation from domestic disagreement, because then any loss can be blamed on outsiders and any dissent can be dismissed as foreign.

What has been tried elsewhere

Bans on foreign money. U.S. law makes it unlawful for a foreign national to make contributions, donations, or expenditures in connection with federal, state, or local elections, or for anyone to solicit or accept them.⁠² Results: the rule is clear. Limits: enforcement depends on tracing money through intermediaries.

Disclosure of foreign agents. The U.S. Foreign Agents Registration Act, enacted in 1938, requires certain agents of foreign principals engaged in political activities to make “periodic public disclosure of their relationship with the foreign principal”.⁠³ It is a disclosure regime, not a ban. Results: it puts relationships on the public record. Limits: it relies on registration and enforcement.

Criminal prohibitions on infiltration. Taiwan’s Anti-Infiltration Act, promulgated January 15, 2020, makes it a crime to make political donations, fund referendum activities, or engage in specified election conduct at the instruction or with the funding of “sources of infiltration,” with penalties up to five years’ imprisonment.⁠⁶ Results and limits: it addresses a specific threat environment, and critics in any democracy would ask how “instruction” is proven without chilling ordinary contact.

Public inquiries. Canada established a Public Inquiry into Foreign Interference in Federal Electoral Processes and Democratic Institutions on September 7, 2023, led by Justice Marie-Josée Hogue, which examined possible interference by China, Russia, and others in the 2019 and 2021 elections and released a final report on January 28, 2025.⁠⁷ Inquiries create a public record; their recommendations depend on later action.

Technical resilience. The 2021 U.S. assessment credits detection through intelligence collection, security monitoring, and “post-election audits” as reasons large-scale manipulation would be hard.⁠¹ Paper records and audits are covered in audits and verification.

Government contact with platforms. Before the 2020 election and 2022 midterms, the FBI and CISA communicated with social media platforms about election misinformation and possible foreign hack-and-leak operations, and CISA forwarded third-party reports to platforms until mid-2022.⁠⁸ In Murthy v. Missouri (June 26, 2024), the Supreme Court held 6 to 3 that the plaintiffs lacked standing and did not reach the First Amendment merits; Justice Alito’s dissent warned, “If a coercive campaign is carried out with enough sophistication, it may get by”.⁠⁸

Competing interpretations

The security-first view holds that foreign influence is a persistent national security threat documented by official assessments,⁠¹ that the cost of under-reacting is high, and that democracies need active monitoring, disclosure laws, and cooperation with platforms to detect covert campaigns.

The civil-liberties view, held by people across the spectrum, holds that the remedy can be worse than the threat when government pressures private platforms to remove speech. Supporters point to the concerns raised in Murthy v. Missouri.⁠⁸ and argue that foreign origin should be exposed through disclosure, while speech by citizens, even if it echoes foreign talking points, must remain protected.

The effect-skeptic view holds that the measured persuasive effect of online influence campaigns is small. Eady and colleagues found no meaningful relationship with attitudes or votes and showed domestic media and politicians dwarfed Russian content in users’ feeds⁠⁴ On this view, overstating foreign power misdirects attention from domestic causes of division.

The trust-erosion view accepts small persuasion effects but argues the real damage is to confidence, which foreign actors target directly with false claims about compromised systems.⁠¹ On this view, transparent audits and clear public communication are the main defenses.

The annulment debate splits people who otherwise agree: protecting an election from a foreign campaign versus preserving the principle that voters, not courts or intelligence agencies, decide.⁠⁵

Application to this framework

My Vote Forward’s design choices respond to several of these risks. Published releases are scheduled, reviewed packages with methods rather than live feeds, which makes it harder to game a public number in real time. The collective priorities chapter covers coordinated participation and why aggregate priorities never carry an automatic mandate. Governance addresses funding independence and who controls the agenda. The identity and registration chapter explains why a future official system would depend on official eligibility, not on account sign-up.

Unresolved questions

• How should the effect of influence campaigns on trust, as distinct from votes, be measured?
• What evidence standard should a court require before annulling an election over an online campaign, and how much of that evidence must be public⁠⁵?
• Where is the line between government warning platforms about foreign operations and government pressuring platforms to suppress citizens’ speech⁠⁸?
• Does disclosure law (FARA style) or prohibition law (anti-infiltration style) better protect elections without chilling ordinary international contact⁠^3,6?