Data ethics

By Luciano Floridi, Mariarosaria Taddeo. 2021

CUSP aims to utilize Big Data to help study and understand urban environments. As a part of this effort, we are planning to build an inclusive data warehouse at CUSP. Our vision for this data warehouse is to hold large quantities of data from multiple sources, including personal (most likely anonymized) data about individuals. But, obtaining, housing, and protecting these data come with many challenges and questions. We hope to answer some of these questions in this working session and converge on a set of principles that will guide our data practices moving forward. Personal data is a new asset class touching all aspects of society. It is potentially as valuable a resource in the 21st century as heavily traded physical goods like oil have been in the past hundred years. However, throughout history, economic value creation has been linked to the ability to move and trade physical goods. Similarly, " data needs to move to create value. Data sitting alone on a server is like money hidden under a mattress. It is safe and secure, but largely stagnant and underutilized. " But, personal data lacks the trading rules and policy frameworks that exist for widely traded physical assets. As a result, there is little trust among the key stakeholders,-individuals, governments and the private sector,-which could undermine its long-term potential. In response to surveys, individuals generally say that they want enhanced control over their personal data, increased transparency on how it is used, and some kind of fair value in return. However, their actions are often quite different. While many say they care deeply about privacy, they share information quite widely online. They often sign up for services not knowing how their data will be protected or whether it will be shared. They rarely read the privacy policies of the organizations providing these services, which are usually written in hard-to-comprehend legal language. Companies, on the other hand, view the data they have captured or created about individuals as theirs. Data is an asset on which they have invested significant resources. They want to leverage the data to create business value, better understand the behavior of their customers and help themselves become more productive. They struggle with how to best protect all the data they now have access to, as well as trying to figure out the different regulations pertaining to its use. Governments are trying to leverage all this data …

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Understanding artificial intelligence ethics and safety (2019, background): “Understanding Artificial Intelligence Ethics and Safety 80 Albarghouthi, A., & Vinitsky, S. (2019). Fairness-aware programming. In Proceedings of the Conference on Fairness, Accountability, and Transparency (pp. 211-219). ACM. Retrieved from https://dl.acm.org /citation.cfm?id=3287588 Chiappa, S., &”
The ethics of AI in health care: A mapping review. (2020, background): “That is, we will focus on the evaluation of moral problems re-lated to data, algorithms and corresponding practices (Floridi and Taddeo, 2016), with the hope of enabling governments and healthcare systems looking to adopt AI-Health to be ethically mindful (Floridi, 2019a).”
Ethics of artificial intelligence and robotics (2020, background): “Opacity is a central issue in what is now often called ‘data ethics’ (Floridi and Taddeo 2016; Mittelstadt and Floridi 2016), particularly in the ethics of big data – whether AI is used there, or not.”
From What to How. An Overview of AI Ethics Tools, Methods and Research to Translate Principles into Practices (2019, background): “…for the development of appropriate governance mechanisms (D’Agostino & Durante, 2018) that will enable society to capitalise on the opportunities, whilst ensuring that human rights are respected (Luciano Floridi & Taddeo, 2016), and fair and ethical decision-making is maintained (Lipton, 2016).”
Artificial Intelligence and Management: The Automation–Augmentation Paradox (2020, background): “Consequently, it is difficult to apportion accountability for outcomes to specific actors, which creates an “ accountability gap ” (Mittelstadt, Allo, Taddeo, Wachter, & Floridi, 2016).”
The ethics of algorithms: key problems and solutions (2020, background): “The interplay between field experts and ML algorithms can prompt “epistemic vices” (Grote and Berens 2020), such as dogmatism or gullibility (Hauer 2019), and hinder the attribution of responsibility in distributed systems (Floridi 2016).”
The promise of machine learning in predicting treatment outcomes in psychiatry (2021, background): “Furthermore, a central issue in fair machine learning develop ment arises when the training dataset is not a good representa tion of the phenomenon being studied 192,208 .”
From What to How: An Initial Review of Publicly Available AI Ethics Tools, Methods and Research to Translate Principles into Practices (2019, background): “…and ‘interpretability’ actually mean when there are currently a myriad definitions (Ananny and Crawford 2018; Bibal and Frénay 2016; Doshi-Velez and Kim 2017; Friedler et al. 2016; Guidotti et al. 2018; Kleinberg et al. 2016; Overdorf et al. 2018; Turilli and Floridi 2009) is standard fare for…”