Simple Demographics Often Identify People Uniquely

By Latanya Sweeney. 2000

Subfields: Privacy & Security, AI Ethics, Fairness & Society

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Privacy & Security

In AI Ethics, Fairness & Society

Citations in the corpus

APPLIED FEDERATED LEARNING: IMPROVING GOOGLE KEYBOARD QUERY SUGGESTIONS (2018, background): “In particular, this reduces the risk of deanonymization via joins with other data [14].”
Communication-Efficient Learning of Deep Networks from Decentralized Data (2016, background): “Holding even an “anonymized” dataset can still put user privacy at risk via joins with other data [37].”
Privacy in Pharmacogenetics: An End-to-End Case Study of Personalized Warfarin Dosing (2014, neutral): “Sweeny showed that a few pieces of identifying information are suitable to identify patients in medical records [42].”
Federated Collaborative Filtering for Privacy-Preserving Personalized Recommendation System (2019, background): “At the servers, the data may be “anonymized” however, anonymization is not foolproof and can still violate user privacy when integrated with other data [2].”
Estimating the success of re-identifications in incomplete datasets using generative models (2019, methodology): “Using socio-demographic, survey, and health datasets, we show that our model exhibits a mean absolute error (MAE) of 0.018 on average in estimating population uniqueness 42 and an MAE of 0.041 in estimating population uniqueness when the model is trained on only a 1% population sample.”
Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures (2015, background): “The classic attack by Sweeney [38] showed that it was possible to correlate publicly-available anonymzied hospital visit data with voter registration data, leading to re-identiﬁcation of some individuals.”
Blockchain Technology for Healthcare: Facilitating the Transition to Patient-Driven Interoperability (2018, background): “First, basic demographic information can identify people [42], and if an individual's public key is matched to their identity, all transactions associated with that public key are then known to be linked to an individual.”
FedNER: Privacy-preserving Medical Named Entity Recognition with Federated Learning (2020, background): “Although different medical platforms may have some own labeled datasets, they cannot be directly aggregated since medical data is highly privacy-sensitive (Sweeney, 2000).”
Observing and Preventing Leakage in MapReduce (2015, background): “Sweeney [25] showed that combinations of such simple demographics often already identify people uniquely.”
Machine learning and genomics: precision medicine versus patient privacy (2018, background): “requires stripping more than unique identiﬁers from the data. As an illustration, consider that an estimated 87% of the U.S. population has a unique combination of date of birth, gender, and zip code [26]. This made it possible to leverage voter registration data to ﬁnd the names of 84 to 97% of the”
FedNER: Medical Named Entity Recognition with Federated Learning (2020, background): “Although different medical platforms may have some own labeled datasets, they cannot be directly aggregated since medical data is highly privacy-sensitive (Sweeney, 2000).”
The psychology of job loss: using social media data to characterize and predict unemployment (2016, background): “Anonymization techniques and data aggregation can be used to protect the users’ privacy, but many researchers have shown that re-identification attacks are a potential threat ([31], [37]).”
Secure Multi-party Differential Privacy (2015, background): “Such attacks have revealed the medical record of a former governor of Massachusetts [37], the purchase history of Amazon users[7], genomic information [25], and movie viewing history of Netflix users [33].”