This session explores issues around accountability, transparency and fairness around our data. By looking at both trends and norms in respecting rights, we will explore the questions around our data and the potentials for it, with respect to redefining how accountability functions and how algorithms are used.

Algorithmic fairness meets the GDPR
Jussi Leppälä, Valmet Corporation

Algorithms make increasingly more significant decisions affecting our lives, and yet algorithmic decisions can sometimes be unfair or discriminatory. Why is algorithmic fairness so important now? Are human decisions actually more biased than algorithmic ones? Is there a commonly accepted definition for algorithmic fairness? Can an algorithm’s neutrality be proven? How can algorithmic fairness be tested? What rights do we have related to algorithmic fairness, human intervention, and algorithmic transparency?

Corporate Accountability and Our Data: Trends and Recommendations
Afef Abrougui, Ranking Digital Rights

The RDR 2018 Index results indicate that users remain largely in the dark about how their information is handled by a company, and have insufficient options to control what is collected and to obtain all the information a company holds on them. This talk will highlight trends in company disclosure and present concrete steps that companies and governments can take to better ensure users’ privacy rights are being respected when companies collect, use, and share our data.

OpenSCHUFA: ourdata & more transparency towards credit-scoring
Walter Palmetshofer

Why we started OpenSCHUFA and why you should care about credit scoring:

Germany’s leading credit rating bureau, SCHUFA, has immense power over people’s lives. A low SCHUFA score means landlords will refuse to rent you an apartment, banks will reject your credit card application and network providers will say ‘computer says no’ to a new Internet contract. But what if your SCHUFA score is low because there are mistakes in your credit history? Or if the score is calculated by a mathematical model that is biased?

The big problem is, we simply don’t know how accurate SCHUFA’s or any other credit scoring data is and how it computes its scores. OpenSCHUFA wants to change this by analyzing thousands of credit records.
This is not just happening in Germany, or just with credit scoring, for example the Chinese government has decided to introduce a scoring system by 2020 that assigns a „social value“ to all residents.

We will present the results from the first 6 month of the campaign, if we reached the goal to reconstruct or approximate the Schufa-Score by crowdsourcing the data.

Session is moderated by: Tuukka Lehtiniemi