This article originally appeared in Law News Day.
By Jordan French
Deepfakes dominate headlines, but election security threats increasingly reflect something bigger: the fusion of cyber operations with influence campaigns.
Since the 2017 Intelligence Community Assessment of Russian interference in the 2016 Presidential election, lawmakers and government agencies alike have publicly warned that foreign actors have manufactured election-related videos intended to undermine trust in U.S. elections and stoke division – an approach that thrives on speed, amplification, and confusion.
That reality shifts the question from “is this video real?” to “how did this narrative get operationalized?” Three pressure points decide outcomes: account integrity, content provenance, and the speed of official response.
“People fixate on the deepfake,” said Elliott Broidy, chairman and CEO of Broidy Capital Holdings. Broidy, whose firm is increasingly focused on technology businesses in defense intelligence, homeland security, public safety, and law enforcement, including AI, argues that the real threat goes deeper. “But the deepfake is just the payload. The real attack is the trust pipeline. Compromise, impersonate, publish, amplify, and then dare institutions to catch up.”
Seen this way, elections face a trust supply chain attack. Adversaries target the systems that make information believable, including identity, distribution channels, and verification, rather than relying on a single convincing fake.
When an official or campaign account is hijacked, a false “polling place closed” post to social media does not need a high-fidelity deepfake to cause harm. It just needs the credibility of a trusted handle. Once a seemingly authoritative source shares it, even sloppy media can spark viral confusion.
Then the mechanics of spread take over. A screenshot outlives a takedown. Minor re-edits can help the same narrative slip past automated matching. Change the caption, crop the video, swap the voiceover, translate it, or re-upload with slight tweaks.
“If an official account gets popped, it doesn’t matter how good your detection tools are,” Broidy said. “You’ve handed the adversary your credibility. And credibility is the one thing you can’t rate-limit.”
That is why the defensive hierarchy is straightforward. Start with identity. Make provenance visible. Treat detection as the backstop. Upstream controls can outperform downstream moderation, but even strong account security does not solve the next problem, validating media quickly when the public is primed to distrust what it sees.
As generative tools improve, purely technical deepfake detection risks becoming an arms race. The National Institute of Standards and Technology (NIST) has described ongoing work evaluating analytic systems against AI-generated deepfakes and the broader challenges of robustness, generalization, and content laundering.
If detection is about spotting deception, provenance is about making authenticity legible fast. One prominent example is the Coalition for Content Provenance and Authenticity, which maintains an open technical standard intended to establish the origin and edit history of digital content through cryptographically signed Content Credentials.
But the last mile is fragile. A video that is downloaded, screen-recorded, or reposted across platforms may lose the metadata that shows where it came from. That can weaken provenance at the exact moment decisions get made.
“Provenance can’t be a niche feature buried in a menu,” Broidy said. “If platforms aren’t preserving authenticity signals end-to-end and surfacing them in the moment, then the public is still flying blind.”
Absent visibility, provenance cannot influence behavior during a crisis, especially for the people who most need it in the first hour. That includes reporters, election officials, campaigns, and community leaders deciding whether to amplify it, deny it, or correct it. People do not share what they have verified. They share what feels urgent and credible.
That is why rapid public communication remains a critical defensive line. In response, federal agencies have issued joint statements attributing fabricated election-related media to foreign influence actors and warning about continued efforts to undermine trust.
“In these incidents, speed beats elegance,” Broidy said. “You need a playbook. Lock the accounts, verify the source, and communicate fast in plain English. If you wait for perfect attribution, the narrative is already baked.”
The most common institutional failure is waiting to respond until everything is confirmed, creating an information vacuum that adversaries fill. The first hour is about stabilizing reality, not winning a debate. The longer uncertainty lingers, the more a single piece of manipulated media becomes evidence of a broader claim, that nothing is knowable.
The broader implication is that election interference has become a trust supply chain attack. Compromise the account, distort the content, amplify the message, and delay the response. The operation can succeed even if the falsehood is later exposed.
Preparing for future election cycles will require more than deepfake detectors. It will require stronger authentication, platform-resilient provenance, and rapid public communication systems that can compete with the time-to-viral clock.
A practical baseline can start small and scale:
- Use phishing-resistant multifactor authentication for all campaign and election-office social and ad accounts
- Limit admin access and monitor login alerts
- Publish originals in one canonical place, retain creation files, and apply Content Credentials where supported
- Maintain a public rumor-control page and a rapid statement protocol with pre-drafted templates
In the AI era, Elliott Broidy and other security leaders emphasize that election resilience will be measured less by perfect detection than by durable identity, visible authenticity, and rapid truth-telling.
About Elliott Broidy
Elliott Broidy is Chairman and CEO of Broidy Capital Holdings, as well as Co-Founder and CEO of LEO Technologies and Chairman of Cynapse.ai.
