Millions will watch the 2028 U.S. presidential race unfold online. However, researchers now spotlight a quieter danger. Coordinated “Election Swarms” may soon saturate digital channels, shaping opinion without detection. These AI-directed collections of agents can mimic human dialogue, amplify fringe narratives, and erode trust in democracy. Moreover, industry telemetry shows malicious bots already outnumber many human users on certain networks. Consequently, security leaders urge rapid countermeasures before campaigns intensify. This article unpacks the threat, pinpoints current evidence, and outlines clear steps for platforms, policymakers, and voters.

Defining Election Swarms Threat

The Science Policy Forum defines an Election Swarm as many AI agents acting toward a common influence goal. Each agent can generate text, schedule posts, and respond contextually across platforms. Furthermore, shared memory lets the swarm learn which messages perform best. Jonas Kunst notes that such coordination “might have disproportionate consequences for democracy.” In contrast, past disinformation used simple bots pushing identical slogans. Swarms instead appear authentic, adopt local slang, and persist for months.

Key properties include persistence, cross-platform mobility, and adaptive learning loops. Moreover, agents can perform A/B testing, quickly discarding content that fails to attract voters. These dynamics produce synthetic consensus, a false appearance of widespread grassroots support. Consequently, policymakers see the tactic as a direct challenge to democratic debate. These characteristics summarize the technical baseline. However, understanding current bot traffic clarifies why scale now matters.

Current Bot Activity Trends

Recent industry reports reveal alarming automation levels. Imperva estimates automated web traffic reached 51% in 2025. Meanwhile, so-called bad bots comprised 37% of that traffic. Microsoft’s Digital Defense team similarly observed rising AI-assisted phishing success. Additionally, Axios found bots driving one-third of social discussion around a Minneapolis incident in January 2026.

• 51% of global web traffic is automated (Imperva 2025).
• 37% of that traffic involves malicious bots.
• 30%+ of selected breaking-news chatter originated from non-human accounts (Axios 2026).

These numbers expose fertile ground for Election Swarms expansion. Moreover, platforms still rely heavily on single-account heuristics, not coordination analytics. Therefore, sophisticated swarms could slip past legacy detection. These realities underscore urgent forecasting for 2028. The next section explores how attackers might escalate.

Predicted 2028 Attack Patterns

Researchers project four main tactics for 2028. Firstly, persona farms will create thousands of durable identities embedded in local communities. Secondly, cross-platform amplification will push aligned messages onto TikTok, X, YouTube, and private chats simultaneously. Thirdly, hijacked real accounts will grant credibility, bypassing bot filters. Finally, synthetic media—text, video, audio—will deepen narrative impact.

Consequently, influence can scale rapidly while remaining personalized. Moreover, swarms could test language variations, then deploy winning phrases in swing districts within hours. Attackers may also contaminate training data, shaping future language models toward partisan angles. These forecasts reveal sharpened manipulation vectors. However, the core societal harm centers on legitimacy.

Core Risks To Democracy

Persistent swarm dialogue may convince undecided voters that false positions are mainstream. Additionally, harassment against journalists could silence fact-checking. Nevertheless, evidence linking bots directly to vote shifts remains sparse. Experts stress that the larger danger lies in eroding shared reality, not ballot tampering. Two critical points emerge. First, misinformation volume can drown authentic voices. Second, doubt about information provenance depresses turnout among confused voters. These concerns frame the needed defenses. Mitigation strategies are advancing, as the next section details.

Mitigation Layers In Development

Platform engineers are building swarm detection dashboards that analyze coordination, timing, and linguistic patterns. Furthermore, provenance standards like C2PA watermarking promise to flag synthetic media. Model developers now run persuasion-risk tests before public releases. Meanwhile, U.S. agencies urge pre-election stress tests and rapid response playbooks.

Professionals can enhance their expertise with the AI Security Level 1 certification. Additionally, the Science authors propose an international “AI Influence Observatory” to share cross-platform telemetry. Consequently, defenders could act on unified signals instead of siloed reports. These layered safeguards show progress. However, success depends on clear policy mandates, explored next.

Policy And Platform Actions

Congressional committees now require transparency reports on coordinated inauthentic behavior. Moreover, CISA, FBI, and ODNI guidance highlights AI threats in election-security planning. Platforms respond by expanding Trust and Safety budgets and offering optional identity verification. Nevertheless, civil-society groups warn about false positives that mislabel legitimate activists as bots. Therefore, balanced oversight remains essential. These evolving rules set the governance backdrop. Practical advice for individual voters follows.

Practical Steps For Voters

Civic participation still hinges on informed citizens. Firstly, voters should cross-check sensational claims with reputable outlets. Secondly, browser extensions that display provenance metadata can flag possible deepfakes. Moreover, pausing before sharing viral content reduces unintentional misinformation spread. Additionally, reporting suspicious coordinated replies helps platforms refine detectors.

Remember that Election Swarms rely on rapid emotional reactions. Consequently, slowing the share cycle blunts their reach. These personal defenses complement institutional safeguards. Yet unanswered questions linger, as the next section explains.

Data Gaps And Uncertainty

Empirical studies rarely measure direct persuasion from bot exposure to ballot choice. Furthermore, platforms limit researcher access to coordination logs, citing privacy concerns. In contrast, independent auditors argue that anonymized datasets could protect users while enabling scientific review. Moreover, the public still lacks confirmation that fully autonomous swarms have operated during large elections. Consequently, analysts debate whether current content quality suffices to alter voter behavior.

Addressing these gaps requires partnerships among academia, industry, and government. Therefore, mandated data sharing and standardized metrics are vital next steps. These unresolved issues inform the closing outlook.

Conclusion And Next Steps

The technical capacity for Election Swarms now exists, and supporting bot infrastructure is widespread. Moreover, 2028 campaigns present an attractive proving ground. Platform detection, provenance technology, and policy oversight are advancing, yet coordination gaps persist. Consequently, stakeholders must integrate layered defenses, rigorous transparency, and public education.

Industry professionals should monitor evolving standards and pursue continuous training. Meanwhile, citizens can adopt simple verification habits that dilute swarm influence. Finally, explore specialized credentials, including the linked AI Security Level 1 certification, to stay ahead of emerging threats.