Chicken and egg all over the place here, as there's no doubt that using feedback from the data can improve the performance of many devices. The issue is whether personal identification can be made anonymous.
Cluster tracing in a pandemic, using mobile phones, eg, is considered an important strategy by many epidemiologists to trace geo location of viral spread in metro areas. This is useful not only for tracking Covid spread, but for tracking spread of other emerging novel viruses, seen inevitable by most virologists.
Anonymous signals used might include geo-correlated mobile search data of symptom queries by those who suspect infection... or correlation, say, with credit card data. Even without contact tracing, one of the behaviors used to flag locations which contributed to the second peak in Covid-19 spread was the use of anonymized credit card data where checks were frequently paid by single-credit cards for a groups dining together, which suggested clustering.
Here's a report of one-such Japanese study, reported in "Nature"....
"Surveillance of early stage COVID-19 clusters using search query logs and mobile device-based location information" Nature - 29 October 2020
[
nature.com...]
