AirTag Stalking: Five Years of Catch-Up on a $29 Tracker

📋 Key Takeaways
  • What happened
  • How it worked
  • Impact and numbers
  • Timeline
  • Why it still matters in 2026
7 min read · 1,266 words
Educational & Ethical Use Only — This article is provided for educational and ethical cybersecurity research purposes only. The techniques described should only be used on systems you own or have explicit permission to test. Always follow responsible disclosure and the laws applicable to you. Mitigations are included so engineers can harden real systems.

When Apple’s AirTag went on sale in April 2021, security researchers immediately flagged what stalking-prevention advocates had predicted: a $29 coin-sized tracker with Apple’s vast Find My network behind it is a perfect stalking tool. Within weeks, real cases surfaced — and the National Network to End Domestic Violence’s Safety Net Project had already published a detection guide before the product even shipped. The following five years produced alerts on phones, Android apps, beeping separators — and an arms race that still hasn’t fully closed a safety gap created by convenience engineering.

Quick Answer
AirTags track items via Bluetooth pings picked up by any nearby Apple device and relayed through the crowdsourced Find My network. That architecture — cheap tracker + network of a billion reporting devices — makes them uniquely effective for covert tracking of people: planted in a car or bag, an AirTag reports its target’s location to its owner. Abuse reports accumulated within weeks of launch (the first police-reported stalking cases by summer 2021). Apple’s mitigations arrived iteratively: alerts to nearby iPhones (iOS 14.5+), an Android scanner app (Tracker Detect, Dec 2021), louder separated-tag beeping logic, a 2023 firmware unknown-tracker alert upgrade, and Precision Finding for located unknown tags. Gaps persist: delayed alerts (8–24h historically), Android-only detection requiring manual app install, and third-party Find My-compatible trackers (Chipolo, Pebblebee, later specification joiners) historically lacked alert parity — closing only partially via the 2024-2025 industry detection-spec harmonisation. Core lesson: safety-by-design must include adversarial-use review for tracking hardware, and detection belongs in the platform, not an optional app.

What happened

The Washington Post’s Geoffrey Fowler documented the risk in May 2021 with a self-experiment (a planted AirTag tracking him), synthesising researcher warnings and NNEDV guidance published in April. Through summer 2021, police reports and court cases surfaced: stalking ex-partners, vehicle tracking for theft, harassment. Prosecutors in Indiana charged the first AirTag-stalking cases by mid-2022; a Rolling Stone compilation in 2022 documented 150+ police reports spanning stalking, harassment, and vehicle-targeting. The pattern was consistent: small, cheap, long-battery (≈1 year), self-reporting via network of bystanders’ devices — a stalking kit for the price of lunch.

Apple’s response timeline became a case study in iterative safety patching: alerts for unknown tags moving with you (iOS 14.5 era), Tracker Detect for Android (December 2021), firmware changes making separated tags beep sooner (2022), a 2023 upgrade cutting unknown-tag alert delay and adding Precision Finding to locate planted tags, and the NPE (noise) adjustments continuing into 2024. Simultaneously, Apple+Google announced a joint specification (2023) for unwanted-tracker detection across platforms — iOS and Android both shipping cross-platform unknown-tracker alerts in 2024, with Chipolo/Moto/Samsung joining. The regulatory layer moved too: multiple state attorneys-general letters, a class-action consolidation alleging AirTag-enabled stalking, and product-safety quiet design changes — all documented across 2022–2025.

The deeper design critique stands: the Find My network’s separation-aware likelihood model was built for owners finding lost keys, not for targets finding adversarial tags. NNEDV and researchers argued from launch that detection defaults (silent network, opt-in scanning on Android, delayed alerts) favoured owner utility over target safety. The 2024–2025 harmonised alerts narrowed that gap only after years of demonstrated harm — the definition of safety debt.

How it worked

The tracker-to-network flow explains both the utility and the abuse:

owner pairs AirTag to Apple ID
tag broadcasts rotating BLE identifier (rotating MAC + payload)
any nearby iPhone/iPad hears it (Find My network)
finder device uploads location (encrypted, owner-only readable)
owner sees tag position on map - in near-real-time

abuse case:
adversary hides tag in target's car/bag/jacket
target carries tag -> tag effectively reports TARGET location
detection asymmetry:
  - iPhone targets: alert (after being near tag for a while)
  - Android targets: nothing unless they install Tracker Detect
  - tag beeps only when separated from owner for hours

The cryptographic design (rotating identifiers, end-to-end encrypted location reports) is genuinely privacy-sound against passive third parties — the threat model it excluded was the carrier of the tag. The exclusion is the instructional core: privacy engineering that protects data-in-transit but ignores physical adversarial use produces polished systems that are still weapons. Modern product-security reviews now include adversarial-abuse (misuse) cases as a first-class track — the AirTag saga is the standard citation, alongside the stalkerware-app ecosystem our identity-threat analysis covers.

data-hmmnm-seam="2">

Impact and numbers

Metric Value Source
Launch 2021-04-30 (Apple Store) Apple
Early risk docs NNEDV guide (Apr 2021); WaPo self-experiment (May 2021) NNEDV/WaPo
Police reports compiled 150+ by mid-2022 Rolling Stone review
Android scanner Tracker Detect app, 2021-12 Google Play release
Alert upgrades 2022 beeping, 2023 shorter delays + Precision Finding Apple firmware/release notes
Cross-platform spec Apple+Google joint announcement 2023-05; shipping 2024 company statements
data-hmmnm-seam="3">

Timeline

Date Event
2021-04 AirTag ships; NNEDV publishes detection guide pre-launch
2021-05–08 WaPo experiment; first stalking reports; researcher teardowns
2021-12 Tracker Detect for Android
2022 Louder/sooner beeps; first prosecutions; 150+ police reports
2023-05 Apple–Google unwanted-tracker spec announced
2024–2025 Cross-platform alerts ship; third-party tracker parity matures
data-hmmnm-seam="4">

Why it still matters in 2026

Trackers are now a permanent consumer category (Apple, Samsung SmartThings Find, Tile, Chipolo, Pebblebee, Moto) — and unwanted tracking detection is a platform-level safety feature users assume exists. The AirTag saga is the canonical teaching case for adversarial design review: it shows the harm timeline (pre-launch predictions → rapid abuse → iterative mitigation over years), the cost of deferred defaults (Android detection gap, alert delays), and how industry standards emerge only after public failure. For security teams, personal- and executive-protection programmes now routinely include sweep-for-tracker protocols (RF detection + platform alerts), applying the same discipline as identity-abuse monitoring. And for product builders, it’s the standing example that “privacy-preserving” cryptography and “safe against misuse” are different axes — both required before hardware ships into the physical world.

data-hmmnm-seam="5">

Detection and hardening takeaways

  • Run platform alerts on every phone you protect. iPhone unknown-tracker alerts are default-on (modern iOS); Android users need to confirm Play Protect scanning and tracker alerts (Settings → Google → Personal). Audit family/executive devices, not just corporate ones.
  • Sweep vehicles and bags physically. Tags hide under bumpers, in wheel wells, inside lining: periodic physical sweep (torch + eyes + RF detector where warranted) catches silent variants and non-networked GPS trackers alike.
  • Teach the beep is unreliable. Separated-tag beeps can be muffled, disabled by re-pairing, or missed in traffic; treat alerts and sweeps as primary, audio as a weak bonus signal.
  • Document and report. Found a tag? Note serial (battery-side), photograph placement, do NOT remove battery before evidence capture if pursuing legal action — then disable (remove battery) and file police report; Apple/Google can be compelled to identify owners via serial.
  • Include trackers in threat models for at-risk staff. DV-survivor guidance doubles for executives/journalists: assume cheap tracking exists, budget countermeasures, and review the identity-attack surface that location exposure feeds.

FAQ

Do AirTags work for tracking people who use Android phones?

Yes — that asymmetry was the core safety gap. Any AirTag near any iPhone reports location; the target’s phone OS was irrelevant to the tag’s reporting, but the target’s detection depended on their platform. Android detection required the manual Tracker Detect app until cross-platform alerts shipped in 2024 — three years after launch. Third-party Find My-compatible tags had uneven alert support until spec harmonisation.

What is the Apple–Google unwanted tracker specification?

A joint 2023-announced standard for how BLE trackers signal and how iOS/Android alert users to unknown tags traveling with them — including third-party tags from participating vendors (Chipolo, eufy/Moto, Samsung, Pebblebee, Tile-lineage). It made detection a platform default rather than a vendor choice, closing the worst gaps by 2024–2025.

Can someone track me with other tools anyway?

Certainly — dedicated GPS trackers, stalkerware apps, and cloud-account compromise predate AirTags. AirTags mattered because they made competent covert tracking cheap, long-battery, and requiring zero technical skill — dropping the barrier changed prevalence, which is why the response had to be systemic (platform alerts + law enforcement familiarity + product redesign) rather than just user advice.

data-hmmnm-seam="end">

Prabhu Kalyan Samal

Application Security Consultant at TCS. Certifications: CompTIA SecurityX, Burp Suite Certified Practitioner, Azure Security Engineer, Azure AI Engineer, Certified Red Team Operator, eWPTX v3, LPT, CompTIA PenTest+, Professional Cloud Security Engineer, SC-900, SC-200, PSPO I, CEH, Oracle Java SE 8, ISP, Six Sigma Green Belt, DELF, AutoCAD. Writing about ethical hacking, security tutorials, and tech education at Hmmnm.