Student Privacy and Big Data

Speaker: Elana Zeide
Date recorded: Jan 21, 2016
The current student privacy regulatory regime does not address the issues raised by modern information technology and data-driven decision-making in education.

Elana Zeide on Student Privacy and Big Data. With the rise of online learning environments, student records are no longer just basic academic and administrative information, but include data and metadata generated from student interaction with digital platforms as well as unexpected sources like student ID badges and social media. Applying big data analytics to this wealth of information has the potential to revolutionize education, but also risks unintended consequences that affect the core values of the education system as well as civil rights and liberties.

The current student privacy regulatory regime does not address the issues raised by modern information technology and data-driven decision-making in education. This presentation highlights key issues of the student privacy debate, proposed reforms, and emerging legal and ethical issues, as well as implications of data-driven education environments and decision-making that extend far beyond school settings.

An AI Pattern Language: Accounting for Human Factors & Human Frames

Speaker: Madeleine Clare Elish
Date recorded: Jan 25, 2017
How are practitioners grappling with the social impacts of AI systems? An AI Pattern Language presents a taxonomy of social challenges that emerged from interviews with a range practitioners working in the intelligent systems and AI industry.

Madeleine Clare Elish presents “An AI Pattern Language,” coauthored with Tim Hwang. The publication is the culmination of two years of research and conversations with a range of industry practitioners working in intelligent systems and artificial intelligence. The work was supported by the John D. and Catherine T. MacArthur Foundation. You can purchase your own copy or download the PDF at autonomy.datasociety.net.

Predictive Policing: Bias In, Bias Out

Speaker: Kristian Lum
Date recorded: Nov 17, 2016
Machine learning algorithms are designed to learn and reproduce patterns in data, but if biased data is used to train these predictive models, the models will reproduce and in some cases amplify those same biases.

Kristian Lum will elaborate on the concept of “bias in, bias out” in machine learning with a simple, non-technical example. She will then demonstrate how applying machine learning to police records can result in the over-policing of historically over-policed communities. Using a case study from Oakland, CA, she will show one specific case of how predictive policing not only perpetuates the biases that were previously encoded in the police data, but – under some circumstances – actually amplifies those biases.

Social Dilemmas Around New Media

Speaker: Ilana Gershon
Date recorded: Mar 17, 2016
We all have moments in which someone’s use of new media baffles us, and we have to ask a friend how to respond. It often isn’t just the content of the message, it is also using that particular medium in that way which leaves us scratching our heads.

Ilana Gershon discusses how we all have moments in which someone’s use of new media baffles us, and we have to ask a friend how to respond. It often isn’t just the content of the message, it is also using that particular medium in that way which leaves us scratching our heads. In this talk, I discuss what anthropological concepts can help us understand our confusion. I will turn to LinkedIn as my case study and analyze the dilemmas people face when using LinkedIn as they look for a job. This will be my starting point to discuss how the newness of new media generates social dilemmas, especially for the people these days who are looking for a job.

Self-regulation in Sensor Society

Speaker: Natasha Schüll
Date recorded: Mar 31, 2016
In the story of self-tracking technology and its increasing automation, a certain ambivalence over the terms of contemporary selfhood comes to the fore.

Natasha Schüll – From the NSA scandal to Facebook’s controversial “mood experiment,” the past decade has seen heated debate over the ways that governments and corporations collect data on citizens and consumers, the ends to which they use it, and the threat this poses to civil liberties. Yet even as this discussion over surveillant monitoring unfolds, the public has embraced practices and products of self-tracking, applying sensor-laden patches, wristbands, and pendants to their own bodies.

Drawing on ethnographic fieldwork, this talk explores how mainstream self-tracking technologies – in their design, marketing, and use – increasingly part ways with the ethos of intensive self-attention found within the Quantified Self (QS) community, serving as digital compasses to guide consumers through the confounding, tempting, and sometimes toxic landscape of everyday choice making and lifestyle management (for instance, by regulating the micro-rhythms of their bites, steps, sips, and breaths). By offering them a way to fulfill the cultural demand for self-management while delegating the often tedious, sometimes existentially taxing labor involved in meeting that demand, such devices at once exemplify and short-circuit ideals of individual agency and responsibility.

In the story of self-tracking technology and its increasing automation, a certain ambivalence over the terms of contemporary selfhood comes to the fore. Are there any connections to be drawn between this ambivalence and broader debates over governmental and corporate surveillance, data privacy, and the possibility for resistance?

Living in a Culture of Algorithms

Speaker: danah boyd
Date recorded: Jun 27, 2016
danah boyd examines the problematic implications of using algorithms designed for one problem to address societal issues without accounting for unintended consequences

danah boyd weaves together her work on youth, privacy, and data-driven technologies, to examine the complicated social and cultural dynamics underpinning social media, the messiness of “big data,” and the problematic implications of using algorithms designed for one problem to address societal issues without accounting for unintended consequences.

The Future of Municipal Open Data, Smart Cities, and Civic Technology

Speaker: Noel Hidalgo
Date recorded: Jul 28, 2016
Lessons learned from working within the City’s civic technology community and designing 21st century systems.

Noel Hidalgo will journey through two fellowships — his Data & Society Fellowship and construction of a new fellowship for 21st century civic hackers. The first half of the discussion will focus on detailed lessons learned from working within the City’s civic technology community, collaborating with CUNY’s Service Corps students, building a municipal open data curriculum, and developing partnerships with the Mayor’s Office, Manhattan Borough President, and various City agencies.

Security and Privacy in a Hyper-connected World

Speaker: Bruce Schneier
Date recorded: Dec 7, 2016
Securing systems where information technology permeates our economies, social interactions, and intimate selves.

Bruce Schneider describes how we’ have created a world where information technology permeates our economies, social interactions, and intimate selves. The combination of mobile, cloud computing, the Internet Things, persistent computing, and autonomy is resulting in something altogether different — a world-sized web. This World-Sized Web promises great benefits, but it is also vulnerable to a host of new threats from users, criminals, corporations, and governments. These threats can now result in physical damage and even death.

In this talk, Schneier will take a retrospective look back at what we’ have learned from past attempts to secure these systems. He will also push us forward to consider seriously what technologies, laws, regulations, economic incentives, and social norms we will need to secure them in the future.

Weapons of Math Destruction

Speaker: Cathy O'Neil
Date recorded: Oct 26, 2016
As algorithms increasingly mediate education, employment, consumer credit, and the criminal justice system, how do we measure their impact on our society?

Tracing her experiences as a mathematician and data scientist working in academia, finance, and advertising, Cathy O’Neil will walk us through what she has learned about the pervasive, opaque, and unaccountable mathematical models that regulate our lives, micromanage our economy, and shape our behavior. Cathy will examine how statistical models often pose as neutral mathematical tools, lending a veneer of objectivity to decisions that can severely harm people at critical life moments.

Cathy will also share her concerns around how these models are trained, optimized, and operated at scale in ways that she deems to be arbitrary and statistically unsound and can lead to pernicious feedback loops that reinforce and magnify inequality in our society, rather than rooting it out. She will also suggest solutions and possibilities for building mathematical models that could lead to greater fairness and less harm and suffering.

Understanding Patterns of Mass Violence with Data and Statistics

Speaker: Patrick Ball
Date recorded: Mar 24, 2016
An exploration how, contrary to the standard assumption, statistical patterns in raw data tend to be quite different than patterns in the world.

Patrick Ball discusses how data about mass violence can seem to offer insights into patterns: is violence getting better, or worse, over time? Is violence directed more against men or women? However, in human rights data collection, we (usually) don’t know what we don’t know — and worse, what we don’t know may be systematically different from what we do know.

This talk will explore the assumption that nearly every project using data must make: that the data are representative of reality in the world. We will explore how, contrary to the standard assumption, statistical patterns in raw data tend to be quite different than patterns in the world. Statistical patterns in data reflect how the data was collected rather than changes in the real-world phenomena data purport to represent.

Using analysis of killings in Iraq, homicides committed by police in the US, killings in the conflict in Syria, and homicides in Colombia, we will contrast patterns in raw data with data in estimated total patterns of violence. The talk will show how biases in raw data can be corrected through estimation, and explain why it matters in these countries, and more generally.

Recorded on 3/24/2016.