Private Dispatch

Given the trend of surveillance capitalism’s exploitation of data, systems which aim to resist this exploitation, must take new precautions to safe guard user data. This of course puts constraints on how digital systems are usually built, especially hard to get right is systems which optimize for efficiency.

Considering that location details are essential to effectively matchmake amongst couriers: “who is best geographically located person for a given job” the difficulty of private dispatching emerges as a worthwhile challenge.

Data Fuzzing

One scheme for allowing co-ordination of two or more parties to collaborate on tasks involving multiple locations whereby the majority of data is encrypted and only essential co-ordination data is unencrypted. All data would need to be cryptographically signed and portable between multiple different systems or software services / apps which will need access. Take for example this snippet from the open orders specification:

{
...
    "date_dispatch": "2020-03-21 12:30",
    "date_delivery": "2020-03-22 17:00",
    "pickup": {},
    "destination": {
        "name": "Anna Hartmann",
        "address": "Other Weg 456, Berlin",
        "availability": "15:00-18:00",
    },
    "suppliers": [
        "ber-ou-1",
        "ber-kugu-1"
    ],
...
}

Simple analysis of the data can determine a significant amount about the transaction happening given that the supplier ber-ou-1 can be understood to represent Original Unverpackt, Wiener Strasse 16 and destination could easily be used to profile a users and their purchasing habits. Consider a “fuzzed” version of the above structure:

    "date_dispatch": "2020-03-21",
    "date_delivery": "2020-03-22",
    "locations": ["KZB-KBT", "KZB-GRZ", "NKN-WSR"],
    "payload": "eyJzaXplIjoibWQiLCJ3ZWlnaHQiOiJhdmVyYWdlIiwic3BlZWQiOiJuZXh0IiwicGlja3VwX25hbWUiOi..."

As you can see the hour of day and pickup, destination.name, address, suppliers values are all hidden encrypted in the payload value and the addition of locations field which lists three IDs KZB-KBT which could be used to represent Kreuzberg, Kottbusser Tor which is useful enough for matchmaking, but privacy respecting enough to resist big data surveillance.