Protecting Society From AI Harms: Amnesty International’s Matt Mahmoudi and Damini Satija (Part One)

**Shervin Khodabandeh** (0:02)
Many of our guests aim to use AI for good in their organizations. On today's episode, we speak with two researchers who focus on protecting human rights when artificial intelligence tools are used.

**Damini Satija** (0:16)
I'm Damini Satija.

**Matt Mahmoudi** (0:17)
And I'm Matt Mahmoudi from Amnesty International.

**Damini Satija** (0:20)
And you're listening to Me, Myself, and AI.

**Sam Ransbotham** (0:23)
Welcome to Me, Myself, and AI, a podcast on artificial intelligence and business. Each episode, we introduce you to someone innovating with AI.
I'm Sam Ransbotham, professor of analytics at Boston College. I'm also the AI and Business Strategy guest editor at MIT Sloan Management Review.

**Shervin Khodabandeh** (0:42)
And I'm Shervin Khodabandeh, senior partner with BCG and one of the leaders of our AI business. Together, MIT SMR and BCG have been researching and publishing on AI since 2017, interviewing hundreds of practitioners and surveying thousands of companies on what it takes to build and to deploy and scale AI capabilities and really transform the way organizations operate.
Welcome. Today, Sam and I are excited to be talking with Matt Mahmoudi and Damini Satija from Amnesty International. Matt, Damini, thanks for joining us today. Let's get started. Matt, tell us a little bit about your role at Amnesty.

**Matt Mahmoudi** (1:26)
Absolutely. And yeah, thanks so much for having us.
I am an advisor and researcher on artificial intelligence and human rights at Amnesty's tech program. My role has been focusing on how certain AI technologies and in particular AI-driven surveillance are taken up by policing agencies, developed by companies, ostensibly for efficiency, but often leading to discriminatory outcomes, inequalities of various forms, and affecting some of the most historically marginalized communities. So over the last couple of years in particular, I've been tracing facial recognition deployments, the companies involved, as well as where police are using the tools. We've looked at facial recognition in places such as New York City, Hyderabad city, and India, and the occupied Palestinian territories, and really paying attention to the ways in which these technologies that promise greater efficiency and promise to be smarter ways of moving people from A to B, or ensuring their safety, are actually leading to the erosion of their rights.

**Sam Ransbotham** (2:29)
Matt, tell us a little bit about what Amnesty International does, what the structure is, how their technology practices got started.

**Matt Mahmoudi** (2:36)
Amnesty International is a movement of over 10 million people worldwide who work together via, for example, volunteering, or through doing research and advocacy and campaigning, to mobilize around key human rights issues of the day, as far as the technology and human rights program is concerned, also known as Amnesty Tech. We're a collective of technologists, researchers, advocates, legal scholars, and more, who work together on trying to hold both companies and states to account on their usage and development of technologies that really put those fundamental human rights at risk. So our work is to investigate and expose the ways in which those configurations of technologies are being used to erode those rights and where possible to advocate for stronger safeguards and regulations and human rights practices that enable us to enjoy those rights even as we continue to live through a rapidly changing world.

**Shervin Khodabandeh** (3:35)
Damini, tell us a little bit about what the Algorithmic Accountability Lab does.

**Damini Satija** (3:40)
Yeah, thank you so much for having us today.
I work in the tech program and I head up a team called the Algorithmic Accountability Lab, a relatively new team within Amnesty Tech. We look specifically at the increasing use of automation and AI technologies in the public sector and within that specifically in welfare contexts and social protection contexts. So we look at how governments and public sector agencies are using automation to determine who gets access to basic essential services like housing, education, health care, benefits, and so on. And our particular interest is in investigating and understanding how these tools have discriminatory impact or disproportionate impact on already marginalized groups, which is something we've already seen evidence of in public sector automation or automation of welfare. And the team itself is a multidisciplinary team of seven individuals, data scientists, human rights researchers, advocacy, legal expertise, a whole range, and to support the vision that we will take a holistic view in interrogating and understanding these systems' impacts on society.

**Shervin Khodabandeh** (4:49)
Thank you for that. This is quite interesting, Sam, because when most of our guests are talking to us about how they use AI to create more profits or more revenues or reduce costs or do good generally, right? It seems like your role is to make sure we don't do bad stuff with AI, right? And so in that context, given your background and expertise in AI, what do you think some of the guiding principles are?