My adolescence, like many of us, was marked by the growing awareness of my privileges (as a Caucasian Swiss man) as well as the numerous injustices rooted in our society. This wake-up call nourished in me the will to be part of the solution, one way or another.
Unfortunately, the routine kicks in, school, university, labor market, and here you are, miles away from the ideals that made you both fearless and purposeful, hiding behind some sterile rationalism. Finally, three years ago, I decided to make the leap and cofounded Gradiant AI to help organizations in the media and cultural industries whose survival is threatened by the development of new technologies, and which, nevertheless, play a crucial role in the stability and social cohesion of our society.
Starting up a business, although often idealized, can be overwhelming and risky, especially when the heart of it happens to be the “naïve” desire to help and make this world a better place.
Thinking human at the core of AI: You are not alone
What a relief when I came across the 700+ members community of DataEthics4All. My first thought was, “I’m not crazy, and if I am, I’m definitely not the only one.”
These 4 days of Bootcamp and hackathon have been very fruitful. On the one hand, the Bootcamp allowed me to deepen my understanding of the ethical issues in data science as well as best practices to mitigate them. On the other hand, the hackathon gave me a real taste of the power and benefits of cognitive diversity in tech while working with a handful of brilliant people from all kinds of backgrounds living all over the world.
Defeating the status quo in the justice system in 48 hours.
For the hackathon, our team chose to tackle the COMPAS case which is an algorithm currently used in the USA in the criminal justice system to assess the reoffense likelihood of defendants. The COMPAS algorithm is harmful to our society since it perpetuates race discrimination in terms of who suffers the harm of incorrect predictions. As an example, African Americans are about twice as likely as Caucasians to be erroneously scored as future recidivists.
In 48 hours, we were able to replicate the COMPAS algorithm while removing race discrimination in terms of who suffers the harm of incorrect predictions. We also developed a web app to help judges make more informed decisions. Unlike the COMPAS algorithm which only displays one single risk score, our app centralizes helpful information and visualizations in a comprehensive way for non-tech people. Visualizing a handful of other information gives more context on how a decision is made by the algorithm, and, by deciphering its « thought » process, demystifies its magical effect that sometimes pushes us to blindly rely on its answers.
In the criminal justice context, where false findings can have far-reaching effects on the lives of people charged with crimes, we believe that AI should not take control of the decision-making but rather be a tool to help judges see cases from different perspectives.
While there is still a lot of work to do and it does not solve the root causes of racial injustices in the criminal justice system, it does prove that, when you bring together a multicultural and multidisciplinary team that genuinely care about social justice, you can challenge the status quo and progress in the right direction, one hackathon at a time.
The Future of AI: A Shared Responsibility
Data science is evolving fast and impact all the aspects of our life. It has already found its way in our daily life, without most of society not even realizing it. Data Science is a tool, and like any tool, it can harm or heal.
As a data scientist and co-founder of a company that tackles AI in a different way, I try to follow ethical guidelines by constantly asking myself about the motivations of the tools I develop and the impact that they can have on people, society, the environment, justice, and life in general. While I believe that, as a practitioner, this individual questioning is essential to prevent AI from harming our society, this is not enough.
That’s why, as a citizen, I also encourage each and every one of us– regardless of our background – to get involved and raise our voices in order to collectively define the space and boundaries of AI practices so those transpire through the legal and institutional frameworks. I thank DataEthics4All for being a key factor in the development of Data and AI literacy.