Gabriela Arriagada Bruneau

Gabriela Arriagada Bruneau

Profile

Supervisors:

Professor Vincent C. Müller – Technical University of Eindhoven (TU/e) – Inter-disciplinary Applied Ethics Centre (IDEA) – The Alan Turing Institute

Professor Mark Gilthorpe – Leeds Institute for Data Analytics (LIDA) – The Alan Turing Institute 

The project:

One of the main discussions in the fields of data science and AI relates to how we can deal with bias and fairness, particularly due to big data’s capacity to reflect societal biases. Biased datasets pose a threat, given that when used to train algorithmic models, they replicate or escalate societal biases, thus making AI systems unfair and discriminatory against minority groups, particularly when used for decision-making processes. Most efforts to overcome unfairness have considered technical de-biasing fixes, merely hinting towards broader ethical implications but without further theorization on how human/ethical notions of bias and fairness influence their technical counterparts.

My research aims to contribute to this debate by developing an ethical framework for data science based on the following analysis:

1. The theoretical distinctions on how to define/understand bias and fairness and their relation. I claim that bias is in a different category from unfairness, and therefore saying 'unfair biased person', 'unfair biased system/algorithm' is invalid under my view. My proposal states that fairness is better understood as a property of state of affairs and bias as a cognitive functionality.

2. The practical and ethical concerns related to discrimination and injustice (inequality, data gaps, and systemic issues), can benefit from the theoretical distinction, helping to address the inequality in the field.

3. Provide guidelines to deal with the elements of fairness and bias that unfold within the field, processes, and outcomes of data science.

I argue that the terminology for bias, in its definition, should reflect the ethical neutrality of the phenomenon of bias, that is, bias is not unfair per se, expanding its scope. Thus challenging the common association of technical and human bias with ethical unfairness in data contexts, and claiming that AI systems, data, or algorithms cannot be unfair or discriminatory in an ethical way. Our moral concerns belong to a different category, that of fairness as justice.

Previous publications

Article:

Arriagada-Bruneau, G., Gilthorpe, M., & Müller, V. C. (2020). The ethical imperatives of the COVID-19 pandemic: A review from data ethics. Veritas46, 13-36. http://dx.doi.org/10.4067/S0718-92732020000200013

Arriagada Bruneau, Gabriela. (2018). Do we have moral obligations towards future people? Addressing the moral vagueness of future environmental scenarios. Veritas, 40, 49-65. https://dx.doi.org/10.4067/S0718-92732018000200049

Book Review:

Denis G. Arnold (ed.) 2009, Ethics and the Business of Biomedicine. Cambridge: Cambridge University Press, in Dilemata Nº 20 (2016), pp. 125-131.

in Spanish: https://www.dilemata.net/revista/index.php/dilemata/article/view/428/419

Remark:

Susan Lufkin Krantz, Refuting Peter Singer’s Ethical Theory: The importance of Human Dignity, Praeger, 155 pp: Westport, in Aporia N° 7 (2014), p. 93-97.

in Spanish: http://ojs.uc.cl/index.php/aporia/issue/view/37

Translation:

David A. Crocker, Enfrentando la desigualdad y la corrupción: Agencia, empoderamiento y desarrollo democrático [Original Title: Confronting inequality and corruption: Agency, empowerment, and democratic development], Veritas Nº34 (2016), pp. 65-76.

in Spanish: https://scielo.conicyt.cl/pdf/veritas/n34/art03.pdf

Other projects

  • THE IDEA POD – Inter-disciplinary Applied Ethics Centre new podcast

I am a presenter at the IDEA pod, a fortnightly podcast that explores and interrogates applied ethics across a range of contemporary issues. Here you will be able to hear interviews on topical concerns for our society and, of course, for us as an Applied Ethics centre. Ranging from fellow postgraduate researchers, master students, to professionals from private and public areas, and a variety of topics such as medical ethics, data and technology, artificial intelligence, philosophy of love, aesthetics, and more.

  • Bringing Philosophy of Science and Technology closer to the Spanish-speaking world

With fellow latin-american PhD students from the University of Cambridge and the University of Bristol, we are preparing an outreach podcast to feature fellow hispanic academic researchers in the discipline and present their work to the public in their native language, in an attempt to expand the access of high-quality research to other communities.

My other web profiles

LinkedIn    Research Gate     Academia     PhilPeople     Twitter

Research interests

Applied Ethics, Data ethics, AI ethics, Ethics and Technology, Bias, Fairness, Equality.

Qualifications

  • MSc in Philosophy - University of Edinburgh
  • BA in Philosophy - Pontifical Catholic University of Chile