The Challenges of Election Security in the Age of AI

Confidence in elections can be undermined without altering a single vote. Creating confusion is sometimes enough to sow doubt. Secure systems may protect votes but still leave voters exposed to manipulation.

In 2024, Iranian cyber actors stole campaign material, attempting to influence the election. U.S. intelligence linked Russian actors to a fake video showing ballots being destroyed in Pennsylvania. These tactics aimed to exploit divisions and erode confidence. With artificial intelligence, this threat becomes more complex.

AI makes deception cheaper and faster. It can produce more convincing fakes easily. However, the risk goes beyond this. AI allows dishonest actors to dismiss real evidence as AI-generated. Social engineering targets human behavior, trust, and decision-making.

During elections, attackers have various tools: emotions, public figures, and constant information. A call mimicking a candidate requesting a donation, a text claiming a polling location change, or a video of an official confessing wrongdoing can mislead. AI can generate such variations on a large scale. Campaign staff can also be targeted. An attacker might impersonate a colleague with a cloned voice to request sensitive information. A single deception could leak private communications, edited to mislead.

Attacks might start with a social media post, continue to a fake website, and finish with a text. Yet, seeing a claim repeatedly does not confirm it. The major issue arises when people doubt everything. Legal scholars Bobby Chesney and Danielle Citron describe this as the “liar’s dividend.” In this scenario, knowledge of deepfakes gives dishonest individuals an excuse to label authentic recordings as AI fakes. Knowing fakes exist shows AI’s capability, not the truth about specific information.

Voters must learn to question suspicious content without assuming all is false. Practical habits make manipulation harder. Firstly, pause when urged to act. Scrutinize urgency and fear. Verify requests through separate channels. Secondly, seek independent confirmation. Multiple accounts repeating a clip may rely on one unverified source. Find credible reporting to establish information sources. Thirdly, contact relevant authorities directly. Verify registration, polling places, and deadlines with state or local election offices. For campaign communication, access the official campaign site instead of unsolicited links. Official statements clarify a campaign’s stance, while independent reporting assesses claims. Lastly, be comfortable saying, “I don’t know yet.” Waiting for evidence ensures no rush to judgment, even when claims align with personal views.

However, responsibility extends beyond voters. Campaigns should secure accounts, maintain accessible original communications, and quickly address impersonation attempts. Platforms need effective mechanisms to report and escalate deceptive election content. Organizations like Defending Digital Campaigns offer cybersecurity resources. AI also plays a defensive role, helping identify malicious infrastructure and coordinated activities to disrupt harmful campaigns early.

Technology alone will not restore trust. As synthetic content proliferates, authentic information must be easier to verify. Voters need trusted sources to confirm official statements and quickly address impersonation incidents. The threat is not just AI-induced false beliefs. Repeated manipulation can lead to total distrust. Election security must address both threats, protecting vote systems and ensuring voters know what is real.