Blind My - An Improved Cryptographic Protocol to Prevent Stalking in Apple's Find My Network
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
In 2020, Apple introduced the Find My protocol, which allows owners to crowdsource the location of their lost Apple devices even when the lost device has no active internet connection (e.g., Wi-Fi, Cellular). The Find My protocol is the basis for Apple's AirTag tracking tokens which were released later in 2021. In order to prevent malicious use of these tokens, Apple also implemented ``item safety alerts'' which can warn a person if they are being tracked by an AirTag without their knowledge. However, researchers have recently identified several shortcomings with these alerts that allow modified AirTags to track unsuspecting victims indefinitely without being detected. Making matters worse, while recognizing the observed malicious use of AirTags, news reports, Apple's press releases, and their intended anti-tracking improvements to the protocol do not consider the potential surreptitious use of the Find My network by custom built AirTag clones. In this work, we present an improved Find My protocol which effectively limits the capabilities of malicious AirTags and guarantees that they can be detected while tracking. We accomplish this by adding additional cryptographic verification into the protocol, which restricts tags to only using a bounded set of keys while tracking. In order to maintain - and exceed - the privacy guarantees of the current Find My protocol, we make use of specialized partial blind signatures. To demonstrate the practicality of this protocol, we implement it end-to-end using a programmable device with the same SoC (nRF52832) as in current AirTags. We also benchmark the cryptographic operations of our protocol and show that they require only modest overhead during the initial pairing procedure.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.004 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.003 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it