D. Fox Harrell, Ruha Benjamin, Behnaz Farahi, Jon Kleinberg

It is now well known that AI systems can encode many forms of bias, including those that serve social oppression. It’s imperative for system designers to consider the impacts of such unfair biases and for users to be critically aware of them. This panel takes up this challenge through an interdisciplinary dialogue to imagine how we can design AI systems to ensure not only ethically sound systems, but systems that serve the needs of human empowerment. Specific topics include: particular forms of bias related to artificial neural network (ANN)-based approaches such as deep learning, social biases such as racism and sexism in AI systems, and the capacity of AI systems to reveal and fight unfair biases — not only to instantiate them.

Co-presented by the MIT Center for Advanced Virtuality and convened by D. Fox Harrell.

D. Fox Harrell, DIGITAL MEDIA AND ARTIFICIAL INTELLIGENCE

Ruha Benjamin, AFRICAN AMERICAN STUDIES

Behnaz Farahi, CREATIVE TECHNOLOGY AND CRITICAL MAKING

Jon Kleinberg, COMPUTER SCIENCE

Bias in AI

Live Presentation and Q&A: Wednesday, April 7, 2021 / 11:00am–1:00pm EST  panel

“Unfolding Intelligence: The Art and Science of Contemporary Computation” gathers artists, scientists, and humanists to discuss aesthetic, technical, and critical issues pertaining to artificial intelligence (AI) and computation. The goal of this interdisciplinary conversation is to bridge popular and tech-world understandings of AI as well as domain-specific, academic, and artistic approaches. The panel discussions stage art-science encounters with the goal of mingling otherwise enclosed areas of research, allowing for new public scrutiny and creating an inclusive field of inquiry that encourages a socially engaged view of our machines.  

The four “Unfolding Intelligence” panels address the following questions: How do recent tools in computation shape the models that scientists, artists, and engineers make of the world and universe? Can artists and scientists create a world in which Traditional Ecological Knowledge (TEK) and artificial intelligence (AI) are meaningfully brought together? Can AI and software systems explain how historically recalcitrant forms of oppression persist, embedded in our technologies? Can these same agents possibly provide alternative ways of being and living together? How has computation shaped the concept of intelligence and what models for the unfolding or formation of ideas does it provide?

Full April 1-9 symposium info: https://unfoldingai.mit.edu