amp.theguardian.com/world/2020/may… Applause to (anonymous members of) this ethics board. You can't appoint an ethics board for rubber stamping. Serious DPIAs may lead to changes in design. Even limited oversight as an afterthought not working out fully here.
Ethics Advisory Board (EAB) letter of 24 April also published, see …a-socialtracking.powerappsportals.com/EAB%20Letter%2…. Great letter, 6 principles.
Response …a-socialtracking.powerappsportals.com/Response%20to%… loosely dated "May", appears to agree to principles.
If I was on the EAB (rather than just EAB by myself), would I be happy?
Response …a-socialtracking.powerappsportals.com/Response%20to%… loosely dated "May", appears to agree to principles.
If I was on the EAB (rather than just EAB by myself), would I be happy?
“The app is intended for 2 purposes”. Odd then that according to the DPIA, the data from the app is intended for 3 purposes: these 2 plus research.
"[... mission creep...] It is because of these concerns that the Ethics Advisory Board (EAB) was established. We hope that your expert guidance on the ethical challenges presented by the technology provides assurance that these concerns will be addressed.”
No, the assurance comes from taking the advice seriously. If mission creep is a main issue as claimed, Gould should have disowned the use for immunity passports and use of data by employers when given opportunities to do so in parliament.
Principle 1: Value “good reason to believe that the app will provide sufficient net-value to citizen/society”.
Response: “Value of the app will be clearly communicated to users”. Leaves out “reason to believe”. Demand of Equalities & Health Inequalities Impact Assessment ignored
Response: “Value of the app will be clearly communicated to users”. Leaves out “reason to believe”. Demand of Equalities & Health Inequalities Impact Assessment ignored
Principle 2: Impact “good reason to believe the app will be effective”. Response leaves out “reason to believe”, no promises of anything that will allow anyone to verify effectiveness.
Principle 3: Security & Privacy.
Response: GDPR security compliance spelled out, consent for additional data, “The data that is donated for research will only be made available to those who have been approved by the NHS“ – rest unconstrained?
Point of proportionality ignored.
Response: GDPR security compliance spelled out, consent for additional data, “The data that is donated for research will only be made available to those who have been approved by the NHS“ – rest unconstrained?
Point of proportionality ignored.
Principle 4: Accountability.
Response: future tense “clear governance structure and engagement path … EAB consultation”.
Suspension or withdrawal of app only linked to effectiveness/impact measurement, no "clear decision points" as asked.
Response: future tense “clear governance structure and engagement path … EAB consultation”.
Suspension or withdrawal of app only linked to effectiveness/impact measurement, no "clear decision points" as asked.
Principle 5: Transparency.
Response: various good points.
“Changes to the app T&Cs will be explained to the user”. No EAB input in that?
NHS will protect data against govt reuse by insisting on court orders where appropriate.
Response: various good points.
“Changes to the app T&Cs will be explained to the user”. No EAB input in that?
NHS will protect data against govt reuse by insisting on court orders where appropriate.
Principle 6: Control
Asked for deletion of all user data including on server when app deleted, and user control over additional functionality.
Both ignored in response.
Asked for deletion of all user data including on server when app deleted, and user control over additional functionality.
Both ignored in response.
So overall: all principles get lip service, some a bit more. Promises to be legal. Users will be patronised. EAB will be consulted. Their requests may be ignored. Decisions will continue to be taken. No, I wouldn't be too happy.
Oh and PS. 4th use of the app documented nowhere except in Ian Levy's blog. 24 hour use of same broadcast id chosen to allow (self-)policing of social distancing.
