General reading impressions - 2020 Week 2

Hello! This week was a bit less productive in terms of reading, both because some deadlines led to more focused interaction with texts and because some of the texts I read in full this week were not that interesting. Still, there were some interesting texts, and I hope my first impressions convey part of what’s good about them.


Alberto Cupani. 2017. Filosofia da Tecnologia: Um Convite [Philosophy of Technology: an invitation]. 3ª edição. Editora UFSC.

A good overview of philosophical technology about technology, presenting the key issues through the analysis of the work of some authors — among them Heidegger, Simondon, Bunge, Ihde, Dreyfus, Feenberg and Winner. I really enjoyed most of the book’s sections, even if the chapter on culture felt a bit too short. Still, I am somewhat bothered by what the book does not cover: it has no treatment of feminist or critical race perspectives on technology, and, despite some discussions in the context of the works of Mario Bunge and Hubert Dreyfus, does not fully engage the impact of the information and communication technologies, focusing on the material technological artifacts and their impacts. Despite those shortcomings, the book is a fine starting point who either had no contact with the philosophy of technology or, coming from the studies of a specific theme, want an overview of the general debates in the field.


Miranda Fricker. 2011 [1988].
Rational Authority and Social Power: Towards a Truly Social Epistemology.
In: Social Epistemology: Essential Readings, edited by Alvin Goldman and Dennis Whitcomb.

Some time ago, professor Liam Kofi Bright mentioned in a Curious Cat answer this collection of essays as a good introduction to Social Epistemology. After too much time, I have followed this recommendation, and the results so far have been good. Since most of the book’s chapters were published as independent papers, I will present separate reviews of those that catch my attention, but an uninitiated reader like me might benefit from Alvin Goldman’s A Guide to Social Epistemology (chapter 3).

Fricker’s paper, presented as the third chapter of this volume, answers both to the traditional epistemologist’s claim that social and political factors are external to knowledge and to the reductionist claim that knowledge consists entirely of such factors. She takes as her starting point Edward Craig’s good-informant account of knowledge, allowing for a distinction between the rational authority (competence and trustworthiness) and the indicator properties in which we usually rely for identifying a good informant. This distinction, in turn, is useful for understanding phenomena such as how the knowledge held by minorities and other vulnerable groups is, in many occasions, disregarded even though it has solid rational bases.

Juliane Kolkott and Christoph Sobotta. 2013.
The distinction between privacy and data protection in the jurisprudence of the CJEU and the ECtHR. International Data Privacy Law 3(4), pp. 223–228.

As the title suggests, this paper attempts to distinguish between two notions that are usually bundled together in discussions about AI: privacy and data protection. While those rights have some overlap, they are not understood as identical, at least within European jurisdictions. For example, the fairness and specific purpose requirements for data protection apply even in cases where privacy intrusion could be justified by other relevant values; not all personal data can be said to be part of an individual’s private life (e.g. criminal convictions only become part of the private life as the event recedes into the past); and the European Court of Human Rights has held that legal persons and non-personified groups can rely on the right to privacy, while data protection is construed as an individual right.

Przemysław Pałka. 2020. Data Management Law for the 2020s: The Lost Origins and the New Needs. Buffalo Law Review 68, forthcoming.

This is an ambitious paper. By rescuing an American report from the 1970s, Pałka intends to sustain some points, each of them interesting on their own:

  1. That the differences between US and EU approaches to data protection law are more superficial than usually thought;
  2. That both approaches to data regulation are insufficient because they are top-down, technocratic, and have an individual orientation;
  3. That the way forward, both for the private and public sectors, requires regulation of the social costs of law, including the collective ones, and fostering public debate about the political choices involved in regulation.

I am highly sympathetic to the author’s thesis, and in previous papers I have argued that socially constructive AI requires the involvement of direct and indirect stakeholders in deciding how AI should be designed and used. Pałka takes that claim one step further, and emphasises the political dimensions of AI, in ways that resonate with Feenberg’s paper that I described last week.

My main issue here is that this approach might be less productive if the political determination of AI is reduced to (slightly improved) representation mechanisms in place in politics. Considering the issues troubling modern democracies — disconnect between representatives and population, concentrations of power, growing extremism, and inertia in key issues such as global warming, among others —, it is healthy to keep in mind that “AI is political” is more of a roadmap to addressing the problem of social cost than a signal of structures that can already be deployed for that.

Thanks to that, the third part of Pałka’s framework — leaving the decisions about AI to political processes —, even if desirable, might face difficulties not just in terms of technocratic efficiency, but also of ensuring actually democratic decisions, especially when it comes to the rights of vulnerable groups. Still, the other two proposals are much more immediately feasible, and point to some low-hanging fruits — such the protection of collective rights, making the society (and not private individuals) the recipient of mandatory notices, or recognizing that different sectors have different regulatory demands — that would greatly mitigate the social costs of AI.

Fabricio Pontin. 2019. Da prudência liberal ao institucionalismo aberto: sobre a necessidade da moderação da liberdade de expressão em Adam Smith, John Stuart-Mill e Amartya Sen. Estudos sobre Amartya Sen 6, pp. 95–106.

Pontin shows how Mill’s defence of the freedom of expression is not as absolutist as usually painted, instead relying on assumptions, similar to those defended by Adam Smith, regarding the moral preconditions for the functioning of markets. Mill and Smith’s positions on the marketplace of ideas, according to Pontin, shared an optimism regarding the role of elites on shaping the space of acceptable discourse, and, given that the last few years allow us to be quite sceptic of such claims, the author suggests that Amartya Sen’s open institutionalism could help address those failures of the marketplace of ideas.

Zupko, Jack. 2018. John Buridan. The Stanford Encyclopedia of Philosophy.

An accessible entry on the work and life of John Buridan, which does a especially great job in showing how his claims about logic are relevant for Buridan’s other works and still present fresh insight for modern debates.

Researcher, Law and Artificial Intelligence

Currently researching the regulation of artificial intelligence at the European University Institute.