Sennay Ghebreab addresses dangers of bias in Artificial Intelligence on Brainwash

12 February 2019

How do we ensure robots don’t make the same mistakes humans do? This is the question Dr. Sennay Ghebreab, AUC’s Head of Studies Social Sciences, recently addressed in an episode of Brainwash. Brainwash (NPO) produces television and radio programmes where philosophers, scientists, writers and comedians challenge audiences by providing them with new ideas and unexpected insights about social issues.

Risks of AI reflecting biased input

In 2015, Google released a photography application that displayed black individuals when the term “gorilla” was searched. Ghebreab uses this example and others from his own lived experiences to illustrate that algorithms are not neutral: “They’re dependent on the input they receive, and on the behaviour of the people that create them,” states Ghebreab.

On the one hand, there exists a serious risk that machines, robots, algorithms and other types of artificial intelligence systems end up copying and propagating negative human behaviour (both consciously and subconsciously). However, Ghebreab also explains that this shortcoming can also be seen as an opportunity for people to learn, as such instances provide insight into various mechanisms of exclusion and lay these hidden biases or tendencies bare. 

Raising awareness through education

According to Ghebreab, not enough companies are aware of the possible dangers that algorithms bring with them, and the ways in which they can propel and exacerbate social and economic inequality. This has motivated Ghebreab to start promoting digital literacy and numeracy, including skills such as being able to use a computer, finding the right information and understanding algorithmic thinking. Ghebreab explains, “These are skills that I teach my students from day one. They are conditions for inclusion and participation in today’s world.”

At AUC, Dr. Ghebreab teaches Information Lab and the theme course Introduction to Information, Communication and Cognition. In his research, he aims to advance our understanding of the human machine by relating computational, behavioural and brain responses to natural pictures and movies. In his courses, he emphasises the importance of digital literacy and numeracy, which is an integral part of contemporary interdisciplinary learning.

Published by  Amsterdam University College