Session: Data Commons and Data Vaults
https://conference.publicspaces.net/en/session/datakluis-datacommonsNaam/name Notulist: Daniël VosDatum/date: 27-06-2023 11:00Sprekersnamen/speaker names:
Shownotes
(mentioned links, books, podcasts, literature, etc.)Live notes (500-750 words)
Martijn van Dam:
Talk about the creation of the personal data vault.
It started with the issue that things on the internet got out of hand, especially behavioural data collectio by 3rd parties.
consequenses of this type of data collection: the company that possess this data have an advantage over parties that do not.
This limits the ability of startups and small companies to start or do business.
Another consequence, this amount of personal data can also be used to influence society and individuals.
Therefore, a new model needs to be developed to regain control.
Whith a personal data vault you get your control back.
by separating personal data from the services.
you can descide which parties have access to your data, using your personal data vault.
advantages:
1. privacy by design
2. equal level playing field for other companies.
What data is collected:
1. Identity data (Social security data)
2. Verified data (auth. data, like a degree)
Most relevant to protect
3. Declared data (preference data)
4. Behavioral data (Everything you click, read, post) <- commercialised the most
datvault will never be used if it does not provide any value, therefore we developed a R & D department to create that value for a vault.
R&D defined 8 research questions ( see powerpoint for these questions)
Cross-sector application ( main value creator for the data vault)
The step-by-step approach to roll out the data vault:
going live in 2 years
Create a testing ground for R&D and startups
Develop interopperability with other data stores
extensive research on governance
start with expanding to other sectors, like culture, retail, ...
CHallenges:
need to be able to scale it.
It only works when sectors together want to use the datavault and cooperate with each other.
Funding, since it requires serious investment to be used in practise.
- Media does not invest, since it does not offer value to them yet.
Ian Forrester
Subject: Personal data stores at the BBC
The BBC want to go digital, but wants to do it in a meaningful way.
BBC's Journey:
Perceptive radio
- Build in 2013, this radio has sensors to tell how close you are, it listens to noise. It collects personal data to decide what it is playing
- "We do not want you data" in collaboration with universities, the BBC came to the conclusion that they do not need your data to operate.
- Databox
- is a project that does not exists anymore, due to technical issues (can download though).
- BBC box project:
- Build on the notion of Human data interaction
- to be able to control the flow of the data within your house
- Not cloud-based, written in GoLang
- Human data interaction
- it is easy to trick users to do something that is bad for you.
- four princiuples:
- 1. Legibility
- 2. Agency
- 3. Negotiability
- 4. Transparancy
My PDS
- Following the databox project, My PDS is based on the solid project.
We were able to pull in data from nextflix and spotify to predict. First it was web-based followed by a UI accesible by phone.
Learnings:
1. There is a larger eco-system in the HDI world
2. Identify frameworks for HDI
3. Change the metric of success, trancending the common metrics like number of clicks.
Panel:
Host: Why approach data as a common?
Marjolein: Applauds the approach of Martijn, but has concerns. Data commons are data sources what they can apply to a commo goal. as a collective you can set the rules for when to share the data. The maintanance of the common is done by the community.
Concerns with the data vault, crusial for the the working of society. The collective needs something different. lack of ethics presentation. We cant pretend that people are rational, people are not aware of what they share. it is a opportunity to educate.
QUestion: how are people going to know?
Martijn: I dont think they lack the knowlegde, yet they still press the accept button, since the other button is to difficult. Privacy is not a reason for people to use a data vault (10-15%), ease of use, and aqcuire value.
Marleen: reacts: Why should people make these descisions and why do companies own these behavioral models. We now need to rebuild our digital world, with public values. We need legislation, ownership.
Compliment Ian and Martijn, that there is a movement to improve the digitalisation.
Ian responds:
Going back to the databox model, you could download a algoritm to the databox. where you could testdrive an algorithm. corperate entities would never freely give the algoritm.
Host: What is the societal value of the data vault
Martijn: Value for individuals as a starting point. for society we have no control on what happens with our data. furthermore, on the note of social media: (1) influences society, (2) they control information, (3) and news intake of society.
Lastly, it helps society to create a more healthy important environment for small businesses.
Marjolijn responds: Commercial part is hard to balance. I heard you saying they can use the data to advertise in a more responsible way. How does this work and how does this help the inidividual
Martijn: Google and meta have the majority of the advertisement market. These commercial parties involved with data vault have an important part of society (newspapers). They have entered a different playing field, since the data they can collect is only 25% owned by dutch companies. what we hope that data vault could influence their bargaining power and improve dutch media.
MArleen responds: I am glad that the economic board is embracing data control now. Data commons is now becoming a serious part forward which is a breakthrough and essential for managing our resources.
Question from the room: It is still about our behavioral data, how do you offer agency for the people using that data vault. What is the responsibility of the data vault?
Martijn: Ethical considerations will be noted in a manifest. and it will be a central part in our governance. Currently, businesses are producing the behavioral data not the people. We want to give control back to individuals, but cannot ask them every questions.
Please use full sentences and write in the spoken language of the session / Graag volledige zinnen gebruiken en de taal aanhouden van de sessie