With AI, Your Entire Internet History is Attributable to you Personally

3 min read Original article ↗

Today, the official account for Senate Republicans posted an AI-generated attack video targeting a Democrat. The video stitches together a series of Reddit posts supposedly made by Graham Platner, the Democrat challenging incumbent Republican Senator Susan Collins for her Maine Senate seat.

X avatar for @NRSC

Senate Republicans@NRSC

Chuck Schumer just coronated Graham Platner as the Democrat nominee in Maine. Meet Graham Platner.

1:19 PM · Apr 30, 2026 · 158K Views

162 Replies · 602 Reposts · 1.33K Likes

The AI video itself isn’t the new development, we’ve seen dramatic, stylized AI videos deployed in politics before. What is largely new is that Platner made these supposed Reddit comments under an anonymous account that was later de-anonymized, reportedly using AI.

AI can now uncover anonymous online accounts. Recent studies show that Large Language Models can analyze writing patterns and cross-reference public data across platforms to identify anonymous users with up to 68% accuracy and 90% precision. The technology makes de-anonymization dramatically faster and cheaper than manual investigation. The consequence are it won’t just be politicians in the future that have their anonymous online activity weaponized against them. Unmasking people and finding the most compromising material can accomplished with a few prompts into your preferred AI.

How AI Uncovers Anonymous Accounts:

  • Pattern Recognition & Writing Style: AI analyzes unique linguistic markers: slang, grammatical habits, and signature phrasing to link anonymous posts to public profiles.

  • Contextual Clues: AI connects small details, such as a mention of a specific pet, a local park, or a “supervisor,” to build a profile of the user.

  • Connecting Digital Traces: AI scans across platforms (Reddit, LinkedIn, “finstas”) to link fragmented information into a recognizable pattern.

  • Analyzing Datasets: AI can identify individuals within supposedly anonymous datasets by evaluating four weeks of mobile phone proximity data.

The implications are massive. The ad tech industry has hinted for years that no one is truly anonymous online, but we have never before seen private citizens specifically unmasked with this technology and then publicly attacked.

This case involves a contentious and consequential Senate race, but there’s no reason to think the technology will stay there. You don’t have to run for office to have your internet history used against you. It could surface when you apply for a job, a loan, or admission to a university. Your views on health insurance might follow you when you try to get coverage for your family.

The theoretical potential for abuse is limited only by imagination.

Going forward we all must assume anything posted online will eventually be attributable to you personally.

As authors of an anonymous blog about contentious political issues, the importance of this subject is not lost on us.

Don’t let this discourage you from voicing your opinion. Don’t self-censor. Instead, be more intentional about what you say. Avoid broad generalizations that can be taken out of context. Don’t pick fights with strangers where no one ends up looking good.

The old adage “say what you mean, and mean what you say” is the best advice going forward. Make sure what you say online is something you would stand by offline.

You will never regret standing up for what you believe in.

Post accordingly.

Discussion about this post

Ready for more?