A scaling law to model the effectiveness of identification techniques
Abstract AI techniques are increasingly being used to identify individuals both offline and online. However, quantifying their effectiveness at scale and, by extension, the risks they pose remains a significant challenge. Here, we propose a two-parameter Bayesian model for exact matching techniques...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Nature Communications |
| Online Access: | https://doi.org/10.1038/s41467-024-55296-6 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841544466567004160 |
|---|---|
| author | Luc Rocher Julien M. Hendrickx Yves-Alexandre de Montjoye |
| author_facet | Luc Rocher Julien M. Hendrickx Yves-Alexandre de Montjoye |
| author_sort | Luc Rocher |
| collection | DOAJ |
| description | Abstract AI techniques are increasingly being used to identify individuals both offline and online. However, quantifying their effectiveness at scale and, by extension, the risks they pose remains a significant challenge. Here, we propose a two-parameter Bayesian model for exact matching techniques and derive an analytical expression for correctness (κ), the fraction of people accurately identified in a population. We then generalize the model to forecast how κ scales from small-scale experiments to the real world, for exact, sparse, and machine learning-based robust identification techniques. Despite having only two degrees of freedom, our method closely fits 476 correctness curves and strongly outperforms curve-fitting methods and entropy-based rules of thumb. Our work provides a principled framework for forecasting the privacy risks posed by identification techniques, while also supporting independent accountability efforts for AI-based biometric systems. |
| format | Article |
| id | doaj-art-cd9c3f8b56e942c29c5f0e3f03cfa3b5 |
| institution | Kabale University |
| issn | 2041-1723 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Nature Communications |
| spelling | doaj-art-cd9c3f8b56e942c29c5f0e3f03cfa3b52025-01-12T12:32:02ZengNature PortfolioNature Communications2041-17232025-01-0116111110.1038/s41467-024-55296-6A scaling law to model the effectiveness of identification techniquesLuc Rocher0Julien M. Hendrickx1Yves-Alexandre de Montjoye2Oxford Internet Institute, University of OxfordInformation and Communication Technologies, Electronics and Applied Mathematics (ICTEAM), Université catholique de LouvainData Science Institute, Imperial College LondonAbstract AI techniques are increasingly being used to identify individuals both offline and online. However, quantifying their effectiveness at scale and, by extension, the risks they pose remains a significant challenge. Here, we propose a two-parameter Bayesian model for exact matching techniques and derive an analytical expression for correctness (κ), the fraction of people accurately identified in a population. We then generalize the model to forecast how κ scales from small-scale experiments to the real world, for exact, sparse, and machine learning-based robust identification techniques. Despite having only two degrees of freedom, our method closely fits 476 correctness curves and strongly outperforms curve-fitting methods and entropy-based rules of thumb. Our work provides a principled framework for forecasting the privacy risks posed by identification techniques, while also supporting independent accountability efforts for AI-based biometric systems.https://doi.org/10.1038/s41467-024-55296-6 |
| spellingShingle | Luc Rocher Julien M. Hendrickx Yves-Alexandre de Montjoye A scaling law to model the effectiveness of identification techniques Nature Communications |
| title | A scaling law to model the effectiveness of identification techniques |
| title_full | A scaling law to model the effectiveness of identification techniques |
| title_fullStr | A scaling law to model the effectiveness of identification techniques |
| title_full_unstemmed | A scaling law to model the effectiveness of identification techniques |
| title_short | A scaling law to model the effectiveness of identification techniques |
| title_sort | scaling law to model the effectiveness of identification techniques |
| url | https://doi.org/10.1038/s41467-024-55296-6 |
| work_keys_str_mv | AT lucrocher ascalinglawtomodeltheeffectivenessofidentificationtechniques AT julienmhendrickx ascalinglawtomodeltheeffectivenessofidentificationtechniques AT yvesalexandredemontjoye ascalinglawtomodeltheeffectivenessofidentificationtechniques AT lucrocher scalinglawtomodeltheeffectivenessofidentificationtechniques AT julienmhendrickx scalinglawtomodeltheeffectivenessofidentificationtechniques AT yvesalexandredemontjoye scalinglawtomodeltheeffectivenessofidentificationtechniques |