Disruption Network Lab

UNTANGLING COMPLEXITY: Working on Anti-Corruption From the International to the Local Level Max Heywood (Transparency International Global Outreach and Advocacy Coordinator, UK/DE) and Stephan Ohme (Lawyer and Financial Expert, Transparency International Deutschland e. V., DE). In the fight against corruption, one of the key elements is making the complex and often hidden system of global economic power transparent. Through chapters in more than 100 countries and an international secretariat in Berlin, Transparency International works on anti-corruption from the international to the local level. In this panel the speakers will address the different ways their organisations are working together on untangling the hidden complexity of the inner mechanisms of the financial system, and present some of the progress that has been made both internationally and locally. ACTIVATION: Collective Strategies to Expose Injustice The Art of Exposing Injustice - Part 4 The 18th conference of the Disruption Network Lab www.disruptionlab.org/activation (https://www.disruptionlab.org/activation) Photo credit: Maria Silvano Produced by Voice Republic For more podcasts visit http://voicerepublic.com

ALGORITHMIC BIAS: AI Traps and Possible Escapes

Caroline Sinders (Machine Learning Designer/User Researcher, Artist, Digital Anthropologist, USA/DE) and Sarah Grant (Media Artist and Educator, Radical Networks, USA/DE) in conversation with Ruth Catlow (Co-Founder and Co-Artistic Director, Furtherfield, UK).

Algorithms are not neutral and unbiased, but instead often reflect, reinforce and automate the current and historical biases and inequalities of society, such as social, racial and gender prejudices. This panel frames this issue, and aims to discuss some possible escapes. Caroline Sinders discusses what an intersectional Feminist AI could look like, and how we could get it. Sarah Grant organises Radical Networks, a community event and arts festival for critical investigations in telecommunications. She will go into how the repeated biases and behaviours that we find in Internet could find themselves patterned and spread into AI systems.

ACTIVATION: Collective Strategies to Expose Injustice

The Art of Exposing Injustice - Part 4

The 18th conference of the Disruption Network Lab www.disruptionlab.org/activation

Photo credit: Maria Silvano

AI TRAPS: Automating Discrimination The Art of Exposing Injustice - Part 2 The 16th conference of the Disruption Network Lab ON THE POLITICS OF AI: Fighting Injustice & Automatic Supremacism Dia Kayyali (Leader of the Tech & Advocacy program at WITNESS, SY/US/DE), Os Keyes (Ada Lovelace Fellow, Human-Centred Design & Engineering, University of Washington, US), Dan McQuillan (Lecturer in Creative & Social Computing at Goldsmiths, University of London, UK). Moderated by Tatiana Bazzichelli (Founding Director, Disruption Network Lab, IT/DE). What do we need to take into consideration when calling for a just AI? Do we need to implement changes in how we design AI, or shall we rather adopt a larger perspective and reflect on how human society is structured? Dia Kayyali will present ways in which AI is facilitating white supremacy, nationalism, racism and transphobia. They will focus on the ways in which AI is being developed and how it is being deployed, in particular for policing and content moderation - two seemingly disparate but politically linked applications. Conversations around AI bias tend to discuss differences in outcome between demographic categories, within notions of race, gender, sexuality, disability or class. Rather than treat these attributes as universal and unquestioned, opening up space for cultural imperialism and presumption in how we "fix" bias, Os Keyes will reflect on how contextually shape these issues. They will discuss how not to fall into the trap of universalise such concepts, arguing that a truly "just" or "equitable" AI must not just be "bias-free" - it must also be local, contextual and meaningfully shaped and controlled by those subject to it. Finally, focusing on machine learning and artificial neural networks, also known as deep learning, Dan McQuillan will reflect on how developing an antifascist AI, influencing our understanding of what is both possible and desirable, and what ought to be. https://www.disruptionlab.org/ai-traps Photo Credit: Maria Silvano Produced by Voice Republic For more podcasts visit http://voicerepublic.com

AI TRAPS: Automating Discrimination The Art of Exposing Injustice - Part 2 The 16th conference of the Disruption Network Lab RACIAL DISCRIMINATION IN THE AGE OF AI: The Future of Civil Rights in the United States Mutale Nkonde (Tech Policy Advisor and Fellow at Data & Society Research Institute, US). Moderated by Rhianna Ilube (Writer, Curator and Host at The Advocacy Academy, UK/DE) To many, the questions posed to Mark Zuckerberg, during the Facebook Congressional Hearings displayed U.S. House and Senate Representatives' lack of technical knowledge. However, legislative officials rely on the expertise of their legislative teams to prepare them for briefings. What the Facebook hearings actually proved were low levels of digital literacy among legislative staffers. Mutale Nkonde will address the epistemological journey taken by congressional staffers about the impact of AI technologies on wider society. Working with low income black communities in New York City, who are fighting the use of facial recognition in public housing, she targets staffers of the Congressional Black Caucus with the goal to advocate for the fair treatment of Black Americans in the United States. She aims to make congressional staffers concerned how police jurisdictions, public housing landlords, retailers, and others have proposed using facial recognition technology as a weapon against African Americans and other people of colour. This talk explores how a conscious understanding of racial bias and AI technology should inform the work of policy makers and the society at large while building up the future of civil rights. https://www.disruptionlab.org/ai-traps Photo Credit: Maria Silvano Produced by Voice Republic For more podcasts visit http://voicerepublic.com

AI TRAPS: Automating Discrimination The Art of Exposing Injustice - Part 2 The 16th conference of the Disruption Network Lab HOW IS GOVERNMENT USING BIG DATA? Crofton Black (Researcher, Journalist & Writer, The Bureau of Investigative Journalism, UK). Moderated by Daniel Eriksson (Head of Technology, Transparency International, SE/DE). AI, algorithms, deep learning, big data - barely a week goes by without a new revelation about our increasingly digital future. Computers will cure cancer, make us richer, prevent crime, decide who gets into the country, determine access to services, map our daily movements, take our jobs away and send us to jail. Successive innovations spark celebration and concern. While new developments offer enticing economic benefits, academics and civil society sound warnings over corporate accountability, the intrusiveness of personal data and the ability of legal frameworks to keep pace with technological challenges. These concerns are particularly acute when it comes to the use of digital technology by governments and the public sector, which are compiling ever larger datasets on citizens as they move towards an increasingly digitised future. Questions abound about what governments are doing with data, who they are paying to do the work, and what the potential outcomes could be, especially for society's most vulnerable people. In May, Crofton Black and Cansu Safak of The Bureau of Investigative Journalism published a report, 'Government Data Systems: The Bureau Investigates', examining what IT systems the UK government has been buying. Their report looks at how to use publicly available data to build a picture of companies, services and projects in this area, through website scraping, automated searches, data analysis and freedom of information requests. In this session Crofton Black will present their work and their findings, and discuss systemic problems of transparency over how the government is spending public money. Report: How is government using big data? The Bureau Investigates(https://www.thebureauinvestigates.com/stories/2019-05-08/algorithms-government-it-systems). https://www.disruptionlab.org/ai-traps Photo Credit: Maria Silvano Produced by Voice Republic For more podcasts visit http://voicerepublic.com

AI TRAPS: Automating Discrimination The Art of Exposing Injustice - Part 2 The 16th conference of the Disruption Network Lab WHAT IS A FEMINIST AI? Possible Feminisms, Possible Internets Charlotte Webb (Co-founder, Feminist Internet & Even Consultancy, UK). Moderated by Lieke Ploeger (Community Director, Disruption Network Lab, NL/DE). What kind of new socio-political imaginary can be instantiated by attempting to design a Feminist Internet? How can feminist methods and values inform the development of less biased technologies? What is a feminist AI? In this keynote, Charlotte Webb will discuss how a collection of artists, designers and creative technologists have been using feminisms, creative practice and technology to explore these questions. She will discuss the challenges of designing a 'Feminist Alexa', which Feminist Internet has been attempting in response to the ways biased voice technologies are saturating markets and colonising homes across the globe. She will discuss how the use of feminist design standards can help ensure that technologies do not knowingly or unknowingly reproduce bias, and introduce the audience to Feminist Internet's most recent project – a feminist chatbot that aims to teach people about AI bias. https://www.disruptionlab.org/ai-traps Photo Credit: Maria Silvano Produced by Voice Republic For more podcasts visit http://voicerepublic.com

AI TRAPS: Automating Discrimination The Art of Exposing Injustice - Part 2 The 16th conference of the Disruption Network Lab AI FOR THE PEOPLE: AI Bias, Ethics & The Common Good Maya Indira Ganesh (Research coordinator, AI & Media Philosophy 'KIM' Research Group, Karlsruhe University of Arts and Design; PhD candidate, Leuphana University, Lüneburg, IN/DE), Slava Jankin (Professor of Data Science and Public Policy at the Hertie School of Governance, UK/DE). Moderated by Nicole Shephard (Researcher on Gender, Technology and Politics of Data, UK/DE). This panel wants to investigate possible solutions to the known challenges related to AI bias, and to discuss the opportunities AI might bring to the public. What can the role of companies, institutions and universities be on how to deal with AI responsibly for the common good? How are public institutions dealing with the ethical usage of AI and what is actually happening on the ground? Maya Indira Ganesh will focus on the seductive idea that we can standardise and manage well-being, goodness, and ethical behaviour in this algorithmically mediated moment. Her talk will examine typologies of policy, computational, industrial, legal, and creative approaches to shaping 'AI ethics' and bias-free algorithms; and critically reflect on the breathless enthusiasm for principles, boards and committees to ensure that AI is ethical. Slava Jankin will reflect on how machine learning can be used for common good in the public sector, focusing on Artificial Intelligence and data science in public services and reflecting on possible products and design implementations. https://www.disruptionlab.org/ai-traps Photo Credit: Maria Silvano Produced by Voice Republic For more podcasts visit http://voicerepublic.com

AI TRAPS: Automating Discrimination The Art of Exposing Injustice - Part 2 The 16th conference of the Disruption Network Lab Adam Harvey (Artist and Researcher, US/DE), Sophie Searcy (Senior Data Scientist at Metis, US). Moderated by Adriana Groh (Head of Program Management, Prototype Fund, DE). Examples abound of AI creating harmful effects and it is good to track and understand these examples. But the next step is harder: how do the practitioners of AI improve what comes next and avoid ever worse and ever bigger "AI catastrophes"? Adam Harvey will present a very concrete art and research project, investigating the ethics, origins, and individual privacy implications of face recognition image datasets and their role in the expansion of biometric surveillance technologies. He will present the latest developments of his project megapixels.cc, showcasing new research on how images posted to Flickr have been used by academic, commercial, and even defense and intelligence agencies around the world for research and development of facial recognition technologies. Drawing on her teaching work on AI ethics and cultural diversity, Sophie Searcy will discuss the fundamental problems underlying the design and implementation of AI and consider how those problems impact to real-world case studies. She will stress out the importance to foster an equal and representative data science community filled with individuals of all technical, educational, and personal backgrounds, understanding the implications of the effect of data science on society at large. https://www.disruptionlab.org/ai-traps Photo Credit: Maria Silvano Produced by Voice Republic For more podcasts visit http://voicerepublic.com

CITIZENS OF EVIDENCE: Independent Investigations for Change The Art of Exposing Injustice - Part 3 The 17th conference of the Disruption Network Lab FORENSIC ARCHITECTURE: Horizontal Verification and the Socialised Production of Evidence Robert Trafford / Forensic Architecture (Researcher, Open Source Investigations, UK). Moderated by Laurie Treffers (Freelance Journalist & Conflict Researcher, Airwars, NL). The institutions on which we have long relied, as producers of truth and to underscore our human rights, are weakened, and our faith in them is declining. New actors, most successfully from the populist right, are inhabiting the space that has opened up. Concepts like "news", "evidence", and "fact" have become weaponised. In this climate, it becomes vital to articulate concepts of truth and truth-making that succeed where present concepts are losing the battle. But the technological and social circumstances that have driven this 'decentralisation' of truth also carry great promise, and propose new possibilities for civil society to expose and interrogate power and its misuse: modern theatres of conflict are densely media-rich environments, information is poorly guarded, and open sources have collapsed the distance between violence and analyst. At the heart of Forensic Architecture's 'counter-forensic' practice is an understanding of the production of evidence as a collaborative enterprise by civil society, in which situated knowledge speaks in coalition with technical, legal, and journalistic expertise. Through the lens of a selection of Forensic Architecture's recent cases, Robert Trafford will discuss how this approach works in practice, its promise, and some of its problems. What are the challenges to traditional forms of truth-production that this 'socialisation' overcomes? What are the avenues through which civil society can expose and prosecute violence, when the state is responsible for that violence? How must this kind of truth be reconstituted when it interacts with legal and political forums? https://www.disruptionlab.org/citizens-of-evidence Photo credit: Elena Veronese Produced by Voice Republic For more podcasts visit http://voicerepublic.com

CITIZENS OF EVIDENCE: Independent Investigations for Change The Art of Exposing Injustice - Part 3 The 17th conference of the Disruption Network Lab TRACKING DICTATORS, DIVING FOR LEAKS & OPENING UP SENSITIVE DATA Emmanuel Freudenthal (Freelance Investigative Journalist, Dictator Alert, FR/KY), Natalie Sedletska (Investigative Reporter & TV Host, UA), Brennan Novak & M C McGrath (Transparency Toolkit, US/DE). Moderated by Shannon Cunningham (Freelance Investigative Journalist, US/DE). This panel focuses on the connection between grassroots investigations and data analysis, and how it is possible to make sensitive data accessible without restriction and open them to the public, facilitating the publication of large datasets. Referring to the collaboration between investigative journalists, "citizen journalists", and the "open-source intelligence" community, Emmanuel Freudenthal will present the methodology of acquiring information behind Dictator Alert – a Twitter bot that tracks the planes of authoritarian regimes installing ADS-B antennas – reflecting on the pros and cons of experimenting with a more "objective" type of information.M C McGrath and Brennan Novak (Transparency Toolkit) will focus on how the tools they are building enable anyone to publish data in searchable archives and sort through large datasets, even without many resources or technical skills, making the publishing of a large sets of documents more accessible to activists and citizen journalists – where previously only large coalitions of media organizations with particularly good technical resources could setup something like this. Furthermore, the panel reflects on the use and impact of open information, and the possibilities for people to contribute a greater variety of insights over time, and the difficultues of relying only on media outlets or single journalists to write the first story, which is often a very hard and complex activity in the case of large datasets. https://www.disruptionlab.org/citizens-of-evidence Photo credit: Elena Veronese Produced by Voice Republic For more podcasts visit http://voicerepublic.com