In this episode of the Campaign Trend Podcast, I talk with John Artunkal, founder and CEO of Argus AI, about what happens when open-source intelligence — the tradecraft built to track sanctions violations and unmask disinformation networks — gets pointed at political candidates. Artunkal trained under Bellingcat, the investigative outfit best known for identifying Russian intelligence officers from selfies and flight manifests, and later studied the social science of the internet at Oxford. Now he applies that tradecraft to opposition research and self-vetting for campaigns.
Much of our conversation centers on digital footprint mapping — building a seed list of a candidate's email addresses, usernames, phone numbers, and even old passwords to map their entire online presence. Artunkal points to the 2024 Mark Robinson case as a textbook example: a reused username, paired with biographical details like hometown and mother's occupation, let researchers cross-reference their way to a full picture. Facial recognition and geolocation tools, borrowed from the OSINT world, are similarly underused in politics, he argues, despite being effective for identifying paid disruptors or verifying who someone really is.
We also dig into what AI changes about the economics of vetting. Campaigns have historically faced a stark choice: no research at all, or a gold-standard human vet costing thousands of dollars few can afford. Artunkal describes the AI-assisted middle tier now emerging — fast and affordable, though imperfect — and the tradeoffs campaigns should understand, including how his platform, Argus X, is tuned to flag borderline cases rather than risk missing a real red flag.
That raises a harder question about accountability. When a human researcher misses something, there's someone to hold responsible. When an AI report comes back clean, there's no way to know what it didn't catch. Artunkal argues responsibility ultimately sits with the tool's designer, and with how clearly they communicate what the tool does and doesn't cover.
The conversation closes on a theme that cuts against conventional campaign caution: candidates shouldn't scrub their online presence into platitudes. Voters respond to authenticity, and a long, consistent posting history actually protects candidates by giving people the full picture of who they are. The real risk isn't imperfection. It's the handful of posts that genuinely contradict who a candidate says they are.