My Thoughts on AI debate and diversity

image credit:

Recent argument by Elon Musk has gained great attention from people working in Machine Learning. Elon Musk worried about the progress of AI research😰. On the other hand, Mark Zuckerberg has positive view on AI🤗. Machine Learning community does not like the Elon’s point of view, Super-intelligence, which is still far away from the current research😤.

So, which opinion is correct?🤔 I think they both miss the point😭. Technology itself would neither improve the quality of life nor threaten society. It is the people who control the technology👀. The biggest concern (at least to me) is diversity rather than AGI (Artificial General Intelligence).

Why diversity in AI matters ? The reason is, ‘Algorithmic bias’😯, as Joy Buolamwini, a student at MIT media lab, tells. On her talk, she talked about the fact that a software did not recognize her face because the training data did not contain the images of black people’s face. She also talked about the algorithm used in judges(e.g., determine how many years a person should spend in prison) does not always result in fair outcome😥. Another problem is the ethics of AI. Iyad Rahwan, a researcher at MIT media lab, talked about this problem using drive-less car example. He and his research found that people think they drive-less cars should be created such that they minimize the total harm. (e.g, saving many pedestrians by swerving you car, which will harm you) but at the same time they do not want to purchase such cars😭. There are so many cases making society be in the loop is important.

I think most of the people working on machine learning are white male and Asian male in Computer Science background (I can’t find statistic behind this).

I think diversity can be divided into 2 parts and both should be addressed correctly:

  1. race: this is related to culture and an important factor for incorporating value, and ethics.👦🏽 👧🏽👦🏼 👧🏼👦🏿 👧🏿👳🏾
  2. background(arts, science, technology): this determines creativities or brings another perspectives

Lowering the entry to this field is really important to solve the diversity issue since the first step into new area is always challenging(e.g, starting programming). One way to do that is communication😍. What I found is that technical people tend to talk with only technical people. Even if you just started studying data science or machine learning, your knowledge is valuable to others outside the field since not many people know how these algorithms work. I think it is ok to make mistakes when trying to explaining the algorithms or democratizing AI . I am very new to this field and make a lot of mistakes😅 By making mistakes, you will learn how to do better just like neural networks do 🤗.

So what will you do or what can we do? Alan Kay said, “The best way to predict the future is to invent it.” Let’s hack the future together😜