Periagoge
Concept
2 min readself knowledge

Fingerprinting: How AI Identifies You Without Cookies or Login

Browser fingerprinting collects dozens of signals—your screen resolution, installed fonts, timezone, JavaScript performance, GPU capabilities—to create a device signature that tracks you across sites even without cookies or login credentials. It's devilishly hard to block completely because fingerprints emerge from legitimate browser data, and advertisers can use even partial fingerprints across weeks or months to track your movement.

Hypatia
Why It Matters

You disable cookies. You use a VPN. You clear your browser history weekly. Yet companies still track you across the internet. How?

Through a technique called browser fingerprinting—a method where AI creates a unique "fingerprint" of your device and browser, which acts like an invisible ID tag that follows you everywhere online.

How Device Fingerprinting Works

Every device sends information when it connects to a website. This information includes your browser type, operating system, screen resolution, installed fonts, timezone, language settings, and your device's capabilities (like GPU specs). Alone, each piece of information is common. But together, the combination is usually unique.

AI uses machine learning to combine these data points into a single fingerprint. The probability that two random devices share the exact same fingerprint is extremely low—often less than 1 in several million. So when your device fingerprint visits a site, then visits another site, AI can confidently connect the dots: same person, same device.

This happens without any tracking code being stored on your device. Nothing to delete. Nothing to block. The fingerprinting happens in the milliseconds when your browser loads a page.

Why This Beats Cookies

Cookies are explicit. A website says, "I'm putting a tracking file on your device." You can delete it, or your browser can block it. Fingerprinting is implicit. It uses information your browser already broadcasts—information required to properly load websites.

It's like someone identifying you by your height, accent, gait, and clothing. You can change your clothing (cookies), but the combination of your other features still identifies you.

The Frustrating Truth

You can't really block fingerprinting the way you can block cookies. Some privacy tools randomize certain signals to make fingerprinting harder (like randomly changing your timezone), but this can break website functionality. VPNs don't help because fingerprinting isn't based on your IP address.

What works somewhat: using a VPN combined with browser tools that randomize fingerprint signals, or using privacy-focused browsers like Brave that actively resist fingerprinting. But these are partial solutions, not complete blockers.

The deeper issue is that fingerprinting is nearly invisible. You have no way to know if a site is fingerprinting you. There's no request, no notification, no permission dialog.

Why Companies Use It

Fingerprinting is valuable because it persists. Cookies can be cleared. Browser profiles can be deleted. But your device fingerprint stays relatively constant for months or years. Companies can track users who deliberately try to avoid tracking—a much more valuable audience for behavioral analysis.

Try this: Visit amiunique.org, which analyzes your browser fingerprint and shows you exactly what data you're broadcasting. See how unique your combination is. Then try the same site with a VPN enabled, or with a privacy-focused browser. You'll see how dramatically fingerprinting can change (or not) with different tools.

Helpful guides
Hypatia
Daily Life & Decisions
Related Concepts
Peri
Questions about Fingerprinting: How AI Identifies You Without Cookies or Login?

Peri can explain this concept, give practical examples, help you decide whether it applies to your situation, or recommend a journey if appropriate.

Ready to work on Fingerprinting: How AI Identifies You Without Cookies or Login?

Explore related journeys or tell Peri what you're working through.