Events

Who the Computer Sees: Race, Gender and AI

Mar 27, 2024

Carla Fehr – University of Waterloo, Wolfe Chair in Scientific and Technological Literacy

Facial recognition systems can do a lot more than open your smartphone. They can sort faces into many categories, including emotional state, age, race, and sex. Most Americans are, without their consent, included in government face recognition databases. This paper develops a case study in which scholar, activist, and public figure Joy Buolamwini diagnoses a now-famous failure of facial recognition systems to ‘see’ and accurately classify Black women’s faces. This case illustrates many important issues in the ethics and politics of AI. In lecture I highlight how this problem is an EDI problem and caution against ‘easy’ solutions that can both backfire and lead to the exploitation of diverse researchers.

Artificial Intelligence, Artifice, and Art

Mar 20, 2024

Ted Chiang – Science Fiction writer (Bio)

Does artificial intelligence deserve to be called intelligence? What are the uses of synthetic text and imagery, and what would it take for those to be artistic mediums?

Disability, Social Media and AI: Implications For the Computing Sciences

Mar 13, 2024

Johnathan Flowers – California State University, Northridge

This talk is divided into two parts: the first part talks through some of the social implications of technology and disability with an emphasis on AI and emerging technologies as they intersect with ableism and ableist discourse in society, covering some of the material in my chapter on AI and disability in the Bloomsbury Guide to Philosophy of Disability. The second part will engage with the computing sciences and disability, specifically the cultural environment of the computing sciences and the ways it relies on a “culture of smartness” which maintains ableism within the field.

The Art of Digital Capitalism

Mar 6, 2024

Tung-Hui Hu – University of Michigan, English & Digital Studies

While the ultimate compliment to an AI model is that it can write poetry or create art, this talk looks to actual writers and artists who have worked alongside digital technology. Moving from the 1970s, when a group of artists decide to build a decentralized network, to the present moment, when artists and writers are training their own AI models, this talk is structured around these footnotes to the history of computation. These artistic works aren’t just decorative or speculative, though; instead, they have the potential to teach us how to live with (and perhaps turn away from) the devastating consequences of digital capitalism.

Pragmatism vs. Principle and How We Get Both Wrong: Inclusive Design Gifts

Feb 14, 2024

Jess Mitchell – Ontario College of Art and Design

From decision-making, thinking, and the creation of everyday things what are we missing? And what hides in the gaps? An inclusive design perspective gives us an opportunity to approach just about everything differently. Let’s have a chat about approaching things differently.

Virtual Reality as Artistic and Reflexive Media

Feb 7, 2024

Sojung Bahng – Queen’s University, Department of Film & Media and DAN School of Drama and Music

This talk will introduce utilization of computational media in artistic and cinematic practices. The primary focus will be on the role of virtual reality as a reflexive material device for storytelling, serving as a lens to reflect our perception and consciousness within socio-cultural contexts.

Algorithmic Bias and Fairness: Exploring Historical Context, Methodological Shortcomings and Future Challenges

Jan 31, 2024

Rina Khan – Queen’s University, School of Computing

AI has seen incredible strides in the past decade and is now ubiquitous in various applications we interact with every day. AI applications have also demonstrated the ability for perpetuating harmful biases and stereotypes to even committing actual harm. This can be observed in facial recognition, law enforcement, hiring screening, automated grading, and natural language processing among others. In this talk, I will examine the lessons that can be learned from the history of computing and AI in relation to algorithmic fairness. I will explore the methodological factors that lead to algorithmic bias and harm and the human factors that are intrinsically interwoven. I will finally discuss proposed mitigation strategies, and the challenges that lie ahead towards creating fairer models.

Designing for Coexistence: Adaptability, Equity, and Ethical Pluralism in Sociotechnical Systems

Jan 17, 2024

Mohammad Rashidujjaman Rifat – University of Toronto, Computer Science

Many technologies today are built on Western scientific principles and empirical data. This approach often overlooks or even discriminates against people whose values are deeply rooted in traditional beliefs and ethics. In his talk, putting faith in the center of analysis, Rifat will examine how the prevailing ethical perspectives in technology development tend to favor some communities while neglecting or marginalizing others worldwide. Rifat will share insights from his diverse research, which spans areas like sustainability, development, privacy, and the prevention of online harms, to highlight how this marginalization occurs. He will then explain his approach to addressing these challenges by integrating theories from postsecular, postcolonial, and decolonial studies with advanced computing techniques, ranging from deep learning to virtual reality. His goal is to develop technologies that are more inclusive, equitable, and plural, especially for those whose ethics and traditions have been overlooked.

Epistemic Corruption and Interested Knowledge

Nov 23, 2023

Sergio Sismondo – Queen’s University, Philosophy

When a system that produces and distributes knowledge importantly loses integrity, ceasing to provide the kinds of trusted knowledge expected of it, we can label this ‘epistemic corruption’. It turns out that such systems are often more fragile than they appear, and they can lose their integrity as a result of internal or external pressures. It also turns out that important actors will often disagree about what constitutes epistemic corruption or which practices are cases – and hence it is important to look at accusations and defences with a measure of neutrality. I will present a small handful examples of epistemic corruption, in an attempt to understand some of the stakes.

Towards Equitable Language Technologies

Nov 16, 2023

Su Lin Blodgett – Microsoft Research Montreal

Language technologies are now ubiquitous. Yet the benefits of these technologies do not accrue evenly to all people, and they can be harmful; they can reproduce stereotypes, prevent speakers of “non-standard” language varieties from participating fully in public discourse, and reinscribe historical patterns of linguistic discrimination. In this talk, I will take a tour through the rapidly emerging body of research examining bias and harm in language technologies and offer some perspective on the many challenges of this work. I will offer some perspective on the many challenges of this work, ranging from how we anticipate and measure language-related harms to how we grapple with the complexities of where and how language technologies are encountered. I will conclude by discussing some future directions towards more equitable technologies.

An Approach Towards Accessibility and Inclusive Design 

Nov 9, 2023

Matt Jacobs and Eric Kellenberger – Queen’s University and San José State University

It is easy to mistake accessibility and inclusive design auxiliary efforts, where accommodations are simply appended to an existing product. In truth, the most extreme examples of need often provide valuable insight into features that benefit everyone. This concept is the driving force behind ‘Universal Design.’ The goal of the present talk is to provide frameworks that broaden our ability to examine accessibility and inclusion, equipping the audience with tools for more iterative, person-driven design approaches. The emphasis will be more on how to interpret and approach the problem, rather than the exact solution itself.

Terraforming Bits & Carbonivorous Clouds: On the Metabolic Rift of Computation

Nov 2, 2023

Steven Gonzalez Monserrate – Goethe University

In the nineteenth century, Karl Marx formulated the concept of “metabolic rift” to describe capitalism’s unsustainable expansion as chemical fertilizers depleted soil nutrients and smog from factories choked the skies of an industrializing Europe. Today, much of what society describes as the “Cloud” resides in data centers not so unlike Marx’s factories. They are the invisible engines of digital capitalism; their pooled, remote computational power and storage capacity are the informatic backbone of everything from social media to payroll to ChatGPT. Like capitalism, computation is a metabolic process. Drawing on six years of ethnographic research in data centers located in the United States, Puerto Rico, Iceland, and Singapore, this lecture surveys the global and local environmental impacts of cloud computing including carbon emissions, water footprint, electronic waste output, land use, and noise pollution. Inspired by science and technology studies and speculative fiction, alternative data ecologies are presented as a corrective to digital capitalism’s environmental excess.

A Critical Look at Canada’s Proposed Artificial Intelligence and Data Act

Oct 26, 2023

Teresa Scassa – University of Ottawa

How do we regulate a technology that crosses all sectors and industries, and that presents considerable risks alongside its promises? Canada’s response to this question, the proposed Artificial Intelligence and Data Act (AIDA), is currently before the INDU committee of Parliament. If passed, AIDA will provide for ex ante regulation of commercial AI in Canada. This presentation offers a critical look at AIDA, placing it within the broader context of other governance work in Canada and abroad.

The Artificial Sublime

Oct 5, 2023

Regina Rini – York University

AI tools like Dall-E, Midjourney, and even ChatGPT can produce objects that look like artwork. But is it really art? Here I will argue that AI is surprisingly well-suited to a particular type of artistic value: the sublime. Sublimity, according to Kant, is the experience of encountering something so vast that the human mind cannot comprehend it. Kant thought that this could be found only in nature, not in art made by humans. But, I will argue, he was wrong about that last part – and it turns out that AI can produce sublime experiences too.

The Future of Work in Canada – A Public Policy Perspective

Sep 28, 2023

Sunil Johal – University of Toronto

Industry Presence and Influence in AI

Sep 7, 2023

Will Aitken – Queen’s University

The advent of transformers, higher computational budgets, and big data has engendered remarkable progress in Natural Language Processing (NLP). Impressive performance of industry pre-trained models has garnered public attention in recent years and made news headlines. That these are industry models is noteworthy. Rarely, if ever, are academic institutes producing exciting new NLP models. Using these models is critical for competing on NLP benchmarks and correspondingly to stay relevant in NLP research. We surveyed 100 papers published at EMNLP 2022 to determine whether this phenomenon constitutes a reliance on industry for NLP publications. We find that there is indeed a substantial reliance. Citations of industry artifacts and contributions across categories is at least three times greater than industry publication rates per year. Quantifying this reliance does not settle how we ought to interpret the results. We discuss two possible perspectives in our discussion: 1) Is collaboration with industry still collaboration in the absence of an alternative? Or 2) has free NLP inquiry been captured by the motivations and research direction of private corporations?

Topic: Copyright and Fair Use

Mar 31, 2023

John Watkinson – Larva Labs

Ethical Issues in the Mass Collection of Human Rights Documentation

Mar 24, 2023

Yvonne Ng – WITNESS

Investigators, researchers, and archivists around the world are using computingtools and services to collect and preserve large quantities of human rightsdocumentary evidence, often without considering all the potential unintendedconsequences and harms. We will discuss some of the ethical issues that arise,and ways that some have found to approach this work responsibly.

Light-touch ethics: Responsible AI’s role in government

Mar 17, 2023

Ana Brandusescu – McGill University

A part of the artificial intelligence (AI) ethics movement, responsible AI has become a dominant strategy in governing AI, rooted in corporate social responsibility. One such example is the algorithmic impact assessment (AIA). Created by governments and professional associations, a typical AIA produces a points-based reward system for impact and risk assessment levels for an AI system. This talk will address power and influence in responsible AI and the broader implications for the governance of AI and its ethics.

Why privacy doesn’t matter and understanding what’s really at stake does.

Mar 3, 2023

LLana James – University of Toronto

Artificial Intelligence: Navigating the Intersection of Ethics, Law, and Policy

Feb 17, 2023

Kassandra McAdams-Roy

This lecture will examine the legal, ethical and policy considerations surrounding the development and use of Artificial Intelligence (AI). The widespread adoption of AI has created new challenges for society, including issues related to data privacy, algorithmic bias, accountability, human safety and more. The lecture will explore some of the current and emerging laws, regulations and other normative frameworks governing AI. It will also discuss the ethical considerations surrounding the use of AI, the broader policy implications, and will consider how best to balance the benefits of this technology with the need to protect individual rights and interests.

What is creativity, and what does it have to do with labour and computers?

Feb 10, 2023

Darren Abramson – Dalhousie University

What does it mean to create something? What do we deserve for our labour? I briefly consider the concept of creativity and argue for a particular view with examples from machine learning. Then I consider the value of labour in programming, and contrast a view from the turn of the millennium with perspectives from recent events.

Inclusive Design, Accessibility and the Outlier Challenge

Feb 3, 2023

Jutta Treviranus – OCADU

What is inclusive design? How is it situated with respect to other forms of design and accessibility? What approaches does it offer for complexity, uncertainty, disparity, and wicked decisions?

Anticheat, Antitox, and the Bottom Line

Jan 27, 2023

Kyle Boerstler – Activision

It seems like a no-brainer that keeping cheaters out of games would be a good idea for companies (and it’s why jobs like mine exist). However, tensions appear when the population of cheaters overlaps with the population of spenders in games. This problem becomes even worse for Antitox, because the perceived harm is lower, often goes unreported, and does not have an obvious solution. For these reasons, investment in antitox is often below investment in anticheat, which means the solutions are often more heavy-handed, and less likely to be implemented because it is even more common for toxic players and spenders to overlap. In this talk, I will cover these issues from my standpoint as a data scientist, addressing the tensions and discussing the relative effects on our player populations.

The Problem with Automated Content Moderation

Jan 20, 2023

Zeerek Talak – Simon Fraser University

Online content moderation using machine learning is a task that is both necessary yet has failed in its mission to protect marginalized communities that are disproportionately at risk of harms. Claims have been made that the issue has been available resources, i.e. datasets and adequately advanced machine learning models. I argue in this talk that the fundamental reason that the power dynamics which govern our social structures have inadequately been subverted to afford the protection of marginalized communities. Through acritical reading of machine learning, I show how the task of protecting marginalized communities is at odds with machine learning without an associated restructuring of the power dynamics that govern the technology.

What Software Eats: The Banal Violences of Efficiency and How to Bite Back

Nov 14, 2022

Bianca Wylie – Digital Public, Tech Reset Canada

Bianca is a writer with a dual background in technology and public engagement. She is a partner at Digital Public and a co-founder of Tech Reset Canada. She worked for several years in the tech sector in operations, infrastructure, corporate training, and product management. Then, as a professional facilitator, she spent several years co-designing, delivering and supporting public consultation processes for various governments and government agencies. She founded the Open Data Institute Toronto in 2014 and co-founded Civic Tech Toronto in 2015.

Bianca’s writing has been published in a range of publications including: Boston Review, VICE, The Globe and Mail, and Toronto Life. She also posts on Medium. She is currently a member of the advisory boards for the Electronic Privacy Information Centre (EPIC), The Computational Democracy Project and the Minderoo Tech & Policy Lab and is a senior fellow at the Centre for International Governance Innovation.

Computing and Global Development: A Critical Perspective

Oct 17, 2022

Ishtiaque Ahmed – University of Toronto

His research interest is situated at the intersection of computer science and the critical social sciences. His work is often motivated by social justice and sustainability issues, and he puts them in the academic contexts of Human-Computer Interaction (HCI) and Information and Communication Technology and Development (ICTD). He operates through a wide range of technical and methodological apparatuses from ethnography to design, and from NLP to tangible user interface.