fbpx
Ki-commerce

4 AI experts who fight algorithmic bias

January 25th 2019
AI Experts
Let's read the text
subscription

Required reading: 5 minutes

"It's not only possible to deal with AI, it's actually critical that you do." With this sentence Rachel Thomas, founder of the organization fast.ai, wants to motivate people with different professional backgrounds to get enthusiastic about artificial intelligence. We introduce you to four AI experts and their projects.


1. Rachel Thomas - founder of Fast.ai, author and spokeswoman

The doctor of mathematics and AI researcher was recently named one of the 20 most incredible women in the AI industry by Forbes magazine. Your company, Fast.aisupports non-profit projects and organizations with knowledge and implementation of artificial intelligence. The expert wants to abolish the exclusivity of AI research and make further education on deep learning and intelligent systems available free of charge. Rachel's enthusiasm for sharing information was inspired by her own background.

"We need more inappropriate people in the A.I. field."

Rachel Thomas


The doctor of mathematics always saw herself as an outsider and was often confronted with sexism, harassment and isolation during her studies in a male domain. Her supervisor even told her that she was too feminine to be successful and stopped the cooperation. As a data scientist and AI researcher, she later switched to the tech industry and experienced this environment as "toxic". The expert sees a high risk in the fact that people - who are responsible for such a toxic working environment - are now developing powerful AI technologies. That's why Rachel Thomas is committed to more diversity in the tech industry.

"We want to make deep learning as accessible as possible - even for users who use uncool languages like C#, uncool operating systems like Windows, uncool datasets (much smaller than Google and in areas you find obscure), and with uncool backgrounds (you may not have gone to Stanford).

Rachel Thomas


It therefore appeals in particular to people who are not confident enough to understand or even implement AI. It doesn't matter whether math is perceived as the biggest weakness or whether the training is second-rate: Not fitting into the picture can be an advantage. Every person has unique knowledge and an individual perspective. These valuable resources can either improve intelligent systems themselves or, in combination with AI, create undreamt-of added value.


2. Joy Boulamwini - computer scientist and activist, founder of the Algorithmic Justice League

The Ghanaian-American computer scientist is on a mission to stop an invisible, growing force. A force, which she herself calls the "coded gaze", also known as the term "algorithmic bias". As a student at Georgia Tech University, Joy Boulamwini developed a social robot to play the cuckoo game known from her childhood. The recognition of the face is an essential part of this posse. But there was a problem: The robot could not recognize its counterpart. The young scholarship holder managed to avoid the challenge by confronting the intelligent machine with her roommate. And I forgot the problem afterwards.

"It matters who programs, how we program, and why we program."

Joy Boulamwini


Years later, the programmer used general facial recognition software to develop an entertaining, intelligent tool. It turned out that the software only reacted when the dark-skinned expert put on a white mask. After a closer look, Boulamwini realized that the same (worldwide) technology was used in both cases. The underlying ML algorithms were trained on the basis of a one-sided data set and therefore never learned what a dark-skinned face looks like. The result: The software could only identify certain skin types and faces contained in the training data set.

"Social change must be a priority, not an afterthought."

Joy Boulamwini


Such algorithms are currently used to identify suspicious persons or to select applicants. Boulamwini recognized the dangerous potential and decided to do something about it. She launched the Algorithmic Justice League - a platform of the MIT Media Lab to identify prejudices by exchanging experiences. Their motto: Everyone can help combat the coded gaze. On codedgaze.com interested people can report prejudices or become testers of intelligent systems. In addition to combating algorithmic bias, the entrepreneur launched several other projects, including the Service Year Initiative to introduce Code4Rights, which helps young people develop meaningful technologies for their communities.


3rd Kate Crawford and Meredith Whittaker - founders of AI Now, researchers at Google and Microsoft

Kate Crawford is living proof that it is possible to enter the field of artificial intelligence at any time. Today's AI researcher started her career in the 1990s as part of the Australian electro music duo "B(if)tek". In 2008, Kate completed her doctorate at the University of Sydney, where she focused on technology research and its impact on the social aspects of our lives. Today she works as a professor at MIT, NYU and University of New South Wales as well as co-founder of various foundations. Her current research focuses on data discrimination and artificial intelligence as well as algorithmic accountability.

"New ethical frameworks for CI must go beyond ownership to account for powerful interests of industry, government and the military in the design and application of CI".

AI Now Institute


In 2016, Kate Crawford worked for Microsoft Research and New York University. Her colleague Meredith Whittaker had a similar position at Google Open Research at the same time. The two researchers joined forces with the White House Office of Science and Technology, the National Economic Council and the Information Law Institute of the NYU. Together they organised the "AI Now" Symposium to examine the future impact of AI. In 2017 this resulted in the AI Now Institutewhich investigates the influence and current technological status of intelligent systems. Research is currently focused on four key areas: Rights and freedoms, work and automation, bias and inclusion, and security and critical infrastructure. Numerous documents, lectures and study results are freely available for viewing.

bottom line

It is both encouraging and fascinating what can arise from knowledge, willpower and passion. Rachel, Joy, Kate and Meredith are undoubtedly pioneers in the fight against discrimination by artificial intelligence. Their work is already bearing fruit and inspired the emergence of similar institutions, such as the Ada Lovelace Institute (named after the founder of programming) in 2018. The individual stories of the four experts prove that AI can and must be accessible and understandable for everyone to achieve change. We can all influence technological development. Intelligent systems are still in their infancy and need to learn a lot. It's up to us to decide what that is.  

Picture: Photo by kinga, EyeEm

These contributions should please you

1 Comment

  • Reply "Enthusiasm for AI can make you blind" - Handelskraft speaker Kathleen Jaedtke and Tina Nord interviewed 31 January 2019 at 10:16

    [...] Technology and what consequences it can have for our society. Keywords: algorithmic bias and bias. Anyway, you never learn [...]

  • Comment now