Concept

Privacy includes whether information reaches the right people under the right conditions.

Issue

Having nothing embarrassing to hide doesn’t answer who needs your information or what they may do with it.

Solutions

Understand the purpose and limits of sharing, then work through one disclosure.

Say a coworker asks why you’ll be away tomorrow. You can explain when you’ll be away without describing your medical history. Yet those medical details may matter when you speak with the doctor treating you.

The data hasn’t changed. The relationship and the reason for sharing have. You can happily tell one person something and still question how another person uses it. Telling someone a piece of data is called a disclosure.

This module is for working adults deciding what to share before they choose to tell someone something.

What you’ll get from this module

In this module

By the end of this module, you’ll have a clearer way to talk about privacy and decide what a disclosure you’re free to make should include. We’ll link a way of thinking about privacy, studies of how people judge sharing, an argument by a law scholar and a real case about health data. You’ll look at what they mean in practice, choose a situation from daily life and work through four questions. You’ll finish with a reason to share, to ask a question first, or to limit or pause something you don’t have to share.

Why the destination and purpose matter

A way of thinking about privacy: information belongs within relationships and conditions.

Helen Nissenbaum’s 2004 account of contextual integrity asks what data fits a particular social setting and how it should move inside or beyond it. This is an argument about what makes sharing appropriate.

A later review by Sebastian Benthall, Seda Gürses and Helen Nissenbaum (2017) explains the parts of this framework. These include who sends data, who receives it, who it is about, what kind of data it is, and the conditions for how it moves.

A setting, or context, means more than an app or a room. It includes people’s roles, what they are doing and the purposes they serve. A colleague acting as a friend may have a different relationship with you from that same colleague arranging cover at work.

Someone knowing a fact for one reason doesn’t on its own justify using it for another. Sharing may be right because of that person’s role or a condition you both agreed. If either changes, the new use needs a reason of its own.

Why it matters: Nissenbaum also allows people to challenge settled expectations when there are good enough reasons. A familiar practice isn’t right just because it’s familiar, and a new use isn’t wrong just because it’s new. The purpose, the interests affected and wider values all matter to that judgement.

Read the original account of contextual integrity. Read the review explaining the framework’s parts.

A study using made-up situations: how sensitive a fact is doesn’t explain people’s judgements.

Kirsten Martin and Helen Nissenbaum (2016) studied 569 people recruited through Mechanical Turk, mostly people in the U.S. Each rated 40 situations put together at random, making 22,760 judgements. The situations changed the type of data, who got it and how it was used. Who sent it and who it was about stayed the same.

Their analysis allowed for each person giving several ratings. In these situations, how data was used or passed on mattered more to people’s sense of privacy than the type of data alone. Who got it and what kind of data it was mattered too.

That finding helps explain why asking whether a fact is sensitive can leave a question half answered. You also need to know who would get it and what would happen next. A judgement about a fact on its own can differ from a judgement about one particular exchange.

Why it matters: The study measured judgements about the situations, not what people actually shared. Its findings are about this mostly U.S. group and the situations used.

Read the study of contextual privacy judgments.

A survey and a check of the rules: steps people accept don’t settle every route.

Yan Shvartzshnaider and colleagues (2016) surveyed more than 450 Mechanical Turk users about how data moves in education. They turned the judgements into rules about allowed sharing, then checked how those rules fitted together.

People approved a teaching assistant sending a struggling student’s attendance record to a professor. They also approved the professor sending it to the head of department. Sending it straight from the assistant to the head got only 17% approval.

The researchers asked whether rules approving each step also allowed a direct route. In this example, they did not. People may have had different ideas about what each role allowed. The result doesn’t show that the people surveyed were being unreasonable.

The route matters as well as the end point. People may accept data passing through a professor while questioning a message sent straight to the head of department. So approving one step doesn’t settle the whole route.

Why it matters: The researchers surveyed people and checked rules; they didn’t test a change in a real school. Fairly few answers per question and the chance that people tired of the survey limited the analysis. You still need to clear up expectations in the real situation.

Read the study of educational sharing expectations.

A law scholar’s argument: a benefit can justify a cost without making the cost disappear.

Daniel J. Solove (2007) looks at the claim that having nothing to hide removes any privacy concern. He considers a stronger version: limited access to certain data could bring a benefit that outweighs a small privacy cost.

His answer separates two claims: there is no privacy problem, or there is a privacy cost but the benefit is worth it. He also looks at concerns about data gathered for one purpose being used for an unrelated one without consent.

That difference gives a conversation somewhere useful to go. You can talk about the actual data, who gets it, the benefit and the limits, instead of arguing about whether someone has a shameful secret. Accepting useful sharing still leaves room to question an unrelated use.

Why it matters: This is an argument about how to weigh privacy against benefits. It helps you frame the question, but it doesn’t decide each case for you.

Read the argument about privacy beyond secrecy.

What decides whether sharing fits?

An official account: the benefit aimed for and the terms of sharing are separate questions.

In July 2017, Dame Fiona Caldicott described how the Royal Free hospital shared data that could identify 1.6 million patients with DeepMind. The work was to build and test an app meant to alert doctors to possible sudden kidney injury.

Caldicott supported new ideas but found that the reason given for the first sharing did not justify it. She separated building and testing technology from giving direct care.

Why it matters: The case shows why a worthwhile aim doesn’t settle whether a particular sharing arrangement is justified. The activity, the data and the conditions still matter. This account from the past describes the benefit aimed for and Caldicott’s view; it doesn’t report measured benefits or harms to patients.

Read the National Data Guardian’s account of the sharing.

Weigh an offer before you share

What this means for you

The useful place to start is one particular disclosure. Knowing why you are sharing and what happens next helps you see what fits, what needs explaining and what is outside your control.

Name the relationship before deciding what fits

Ask what the exchange is for, who is getting the data and whose data it is. Include someone else’s interests when the data is also about them. Your own comfort with sharing isn’t the only thing to weigh.

In the made-up coworker example, arranging cover may need to know when you’ll be away. The relevant medical details serve a different purpose in a talk with the doctor treating you. How much is right depends on what the person needs for that task.

If the coworker is also a friend, name which relationship matters in this conversation. You might be glad to talk about something personal as friends while wanting it kept out of a wider work discussion. Make that expectation clear instead of assuming you both see the same line.

Don’t hold back facts that are essential to the right care just to keep a disclosure small. The aim is relevant sharing under the right conditions.

Separate assumptions from commitments

Ask how the person expects to use, keep or pass on the data. Tell apart what someone has clearly agreed to from what you assume because the relationship feels familiar.

Then work out who else may get the data and how they may use it. The education survey shows why one accepted step doesn’t settle a whole chain. Agreeing to tell one person need not answer what someone in another role may do with it.

Not knowing how data may be passed on doesn’t prove anyone has done wrong. If the answer could change your decision, ask before you share by choice. A clear answer can help you choose, although a promise can’t control what happens later.

Weigh the particular benefit and the wider use

Think about what sharing makes possible and who benefits. Compare that benefit with the extra data, people and consequences involved. A new use needs a reason that holds up, and being new doesn’t rule it out.

Solove’s distinction helps here: admit the privacy concern before deciding whether the benefit outweighs it. The Royal Free account adds a real reminder that you can support an aim and still question a sharing arrangement.

You can put it simply: you may have nothing embarrassing to hide and still care who uses your data. If someone suggests a disclosure that would help, look at that particular proposal together. Neither unease nor a claimed benefit settles the comparison on its own.

Choose an action within your control

Share relevant data when the person needs it for the stated purpose and the conditions are right. Ask a question first when the answer could change your choice. Limit or pause something you don’t have to share when an unneeded use is still unanswered.

These choices affect what you share now. They can’t promise what organisations or later recipients will do. Disclosures you must make, or that are essential, may leave you with a concern and no real alternative; that doesn’t make later misuse your fault.

So a useful result can be a clear question. You don’t need to force an unclear situation into a confident yes or no.

Who should hear this from you?

How about your daily life?

Where do you face a choice about explaining an absence, sharing a personal update or passing on data that’s about someone else? Think about something you don’t have to share whose purpose or later use you’d like to understand.

Write one or more possibilities. One suitable idea is enough.

A situation in my daily lifeWhat I want to clarify
________________________________________
________________________________________

My chosen situation: ____________________

Choose one situation to carry into the exercise below. Choosing it means you’ll look at it; it doesn’t yet show that sharing or holding back is right.

Work through one disclosure

Start with your chosen situation. If you came straight here, choose a disclosure you’re free to make soon, with consequences worth thinking about. Describe the situation without having to write down the personal data itself.

Don’t hold back facts essential to the right care for the sake of this exercise. If the disclosure is required or essential, write down the concern and the options you have, without assuming you can refuse.

  1. What is this for? Name the activity, the role of the person getting it and whose data it is. Include anyone else whose interests the disclosure touches.
    Purpose, role and person concerned: ____________________
  2. What fits that purpose? Work out what the person needs. Write down expectations about using, keeping or passing it on, keeping clear commitments apart from assumptions.
    Relevant information and conditions: ____________________
  3. Where else may it go? Work out who else may get the data and how they may use it. Mark what you don’t know; not knowing doesn’t on its own show anyone has done wrong.
    Known onward uses and unanswered questions: ____________________
  4. Would that wider use be justified? Compare the purpose and benefit with the extra data, people and consequences. Think about the reason for a new use rather than rejecting it just because it’s new.
    Reason the use fits, or concern that remains: ____________________

Use your answers to choose what to do next. Share relevant data when the person needs it for the stated purpose and the conditions are right. Ask a question when an answer that matters is missing. Limit or pause something you don’t have to share when an unneeded use is still unanswered.

My next action, question or unresolved limit: ____________________

Ask Jing

Become a Jingius.

Think one disclosure through

Loading interactive tool…

Your next step

You have finished when you can explain your choice and what is still unclear. This record helps with the decision now; it can’t make anyone keep a promise or control later uses. Organisations stay responsible for what they do.

A note on the evidence

The framework and the law scholar’s argument explain how to judge sharing. The surveys describe judgements in particular settings, including disagreements that written-out rules can bring to light. The past case shows a disputed sharing arrangement alongside a benefit that was aimed for.

People can disagree about roles and purposes. Treat that disagreement as something to clear up, while keeping apart a personal decision and control over what others do later.

How we prepared this module

This module links privacy scholarship, studies of how people judge sharing in context and an official account of a past case. A made-up coworker example makes the difference between roles concrete. The written exercise applies the ideas to one disclosure you’re free to make.

Sources