How do intersectional identities interact with data systems?
A talk on behalf of the Ada Lovelace Institute.
To listen to the talk and see subtitles, see this link> https://otter.ai/u/gF8-NBSRhB9nl2r9ah93YA5c2lc
2. Why inclusive
approaches to
data matter –
The feedback loop
engendered by effective
inclusive approaches
As set out in a forthcoming Ada Lovelace
Institute report – ‘Exploring Participatory
Mechanisms for Data Stewardship’ (2021)
3. From that timeless classic…..
“Who are you?” said the Caterpillar.
This was not an encouraging opening for a
conversation.
Alice replied, rather shyly, “I—I hardly
know, Sir, just at present—at least I know
who I was when I got up this morning, but I
think I must have been changed several
times since then.”
“What do you mean by that?” said the
Caterpillar, sternly. “Explain yourself!”
“I can’t explain myself, I’m afraid, Sir,” said
Alice, “because I am not myself, you see.”
4. ‘the interconnected nature
of social categorisations
such as race, class, and
gender as they apply to a
given individual or group,
regarded as creating
overlapping and
interdependent systems of
discrimination or
disadvantage.’
A (working) definition of
intersectionality
5. Some reflections on this definition
- The role of privilege as well as disadvantage>
Intersectional identities can privilege as well as disadvantage or oppress. You can
simultaneously experience privilege and oppression. But you are more likely to be aware
of your own disadvantage than of privilege.
- Identities, not identity>
Not always fixed, dynamic relations across themselves, has potential to be fluid, change
over time, be socially constructed or ‘performative’ – remember Carroll’s ‘who are you?’
question
- The invisibility question and the right to refusal>
Not everyone’s experience of privilege and oppression is visible – invisibility and the right
to remain invisible or the right to ‘refusal’ of data systems (without repercussion or
coercion) is at the heart of the debate about inclusive data). Should we expect everyone to
‘reveal themselves’? Are there less intrusive ways of getting at the issues?
- Visibility can be stigmatising>
In the case of disadvantaging characteristics – these can be (and historically) have had the
effect of being stigmatising and sometimes criminalising – which can contribute to
legitimate concern about misuse of data in an untrustworthy ecosystem. Think of eugenics,
the Holocaust, mental health stigma, state responses to HIV/AIDS, the list is endless…
6. But identity is one part of the picture. No
person is an island…..
Dahlgren-Whitehead rainbow, 1991
7. Issues raised by the Shielded Patients
List/QCovid algorithm
- Inclusion can compound
inequalities
- Exclusion can compound
inequalities
- Difficult decision making
here, requiring careful
balance and thoughtful,
considered approach to risk
mitigation and design/roll
out
- Inclusion, dialogue with
affected communities and
impact assessment vital
Editor's Notes
There is a feedback loop between getting data right – in ways that work fairly for everyone, and how people and different perspectives and communities feel about data systems. Inclusive data is a symptom – we’ll know we’ll have got things right when we don’t see inequities play out at that level, but it’s not a simple solution or panacea. We can’t just ‘fill in the gaps’ – it’s more complicated than that.
The challenge for inclusive data is captured in a wonderful Lewis Carroll dialogue between a caterpillar and Alice in Wonderland. The caterpillar asks ‘who are you’. In the context of the book of course, Alice has gone through an extraordinary journey and met all sorts of things and people, encountered all sorts of barriers and has fundamentally evolved and changed through the process. Thus the challenge for datafication – because we’ve classified someone as ‘something’ at a given point in time, does not mean that this is accurate or reflective, or even reflects Alice’s own conception of herself. There are limits to inclusive data.
The key point here is that we are as much shaped and conditioned by our environment as we have ‘identities’ within ourselves. Back again to the Lewis Carroll quote. So we need to understand intersectional identities but we also need to understand that design of the environment (socio technical systems) can influence and shape identities too. In ways that are challenging and problematic. Facial Recognition Technology systems that classify people’s gender. Or FRT that classify people’s ethnicities. No choice, no control over ability to define identity.