STS-3094B Diversity and Technology

First off, before this class and outside of this class I didn't know anything about the NYU aging incubator until today. I learned I liked a lot of programs such as the student photo competition and the buddy program. I also think that it is really good these programs have continued in a modified fashion throughout quarantine and the COVID-19 pandemic. Overall I think they have a great mission. It is good to remember that aging and getting old isn't a bad thing. Western culture is strongly and oddly against aging, and -- although they aren't married ideas -- death, as well. In a lot of other cultures death is considered just another part of life, and aging is a sign that somebody should be respected. Just recently, my grandmother passed, and the mayor of her town came and organized the flowers at her wake personally. I think this goes to show that aging should not be stigmatized as much as it is. The way we express our disdain for aging is very subtle, but it is still there...

The other midterm project that I looked into was Helen's about voice controlled devices and language as a salient social identity. Helen and I were partners in this class at the beginning of the semester. We both chose to target language as the salient social identity we were going to grapple with. I ended up choosing natural language processing as my technology, and I believe that our projects therefore end up complementing each other. Our projects were very similar in the issues that were brought up, however Helen approached language from the lens of accents and I approached language from the lens of vernacular. I think that if natural language processing were to consider all vernaculars and voice control devices were to consider all accents, we would have seamless voice controlled devices that are accessible to all. Making this type of human to computer communication accessible to all would remove a lot of barriers in the world of technology 

Presentation Problem: Natural language processing is generally considered to have started with Alan Turing’s work in the 1950s, although the implementation of artificial intelligence into the field in the last decade has given its developers a social responsibility in preventing their technology from perpetuating demographic biases.  This includes sexual and racial bias, as well as language bias. While the technology is researched worldwide, the diversity in the United States makes it especially important for American developers to be vigilant with these issues. English may be spoken by 1.75 million globally (Harvard Business Review), but vernacular and skill level brings up the question, what can be considered natural language? Affected Populations, SSIs: “In psychology, where most studies are based on western, educated, industrialized, rich, and democratic re- search participants (so-called WEIRD, Henrich et al. (2010)), the tacit assumption that human nat...

Algorithms of Oppression , while dated, demonstrates Google's disregard for bias prevention.  My midterm project is on how NLP can open up perpetuation of demographic biases, and the examples drawn from the book are directly related to this.  Today's discrimination can take on a more insidious form, less and less lunging to tear down and more and more hiding behind systemizing the us/them mentality.  This is the type of discrimination brought up in Algorithms of Oppression .  Google never outright states that women of color are sexual objects but is certainly ready to direct you to HotBlackPussy.com upon a keyword search "black girls" (Safiya 3). One of the Google Talk's proposed solutions that came up fairly often was diversifying their team.  While this is a good idea generally, it brought up a lot of notions that did not sit well with me.  The most obvious of which is the fact that the need to diversify your team indicates there isn't already...

I have always been passionate about the role language plays in a person's life.  In my family, I was the first person to be born in the United States, and my parents made sure I was raised "All-American."  As they learned about American culture, I was raised in it.  I wasn't taught to cook adobo or speak Cebuano or stay in touch with my heritage.  My parents would always speak to me in Cebuano, but always expected me to reply in English.  I grew up with the ability to understand what my parents told me, but without the ability to make them understand everything I told them.  Much of my life was underscored by this jarring disconnect. Eventually, I was able to work my way into broken but understandable Cebuano, but even still I find language to be one of the most interesting and abstract ideas I have put thought into.  My senior English project in high school addressed language barriers, and I wanted to bring back a lot of the powerful se...

In the Harvard case, lacking context makes it hard to determine who is right.  However, as you go deeper into the case, it becomes increasingly clear there is no hard evidence in the plaintiff's favor.  What this case seems like is an instance of faulty causality, where the lower population of Asian students at Harvard is framed as a lower population of students because they are Asian.  While it is easy to complain that there were qualified Asian students with high GPAs and test scores, there is an admissions process for a reason - you won't get into the school just because you meet a benchmark.  Furthermore, looking at the case from this perspective, it can feel as though the accusation is naive to factors outside of intelligence that impact school performance.  Coming back to an old source ( Questioning the 3D Printing Revolution ) from last blog post, " marginalized people have always created goods, products, arts, platforms and businesses,...

Updates #1: Possible ways to address diversity in the field of natural language processing (NLP): Implementing an Institutional Review Board (IRB) has not yet been considered in the field of NLP, but is a generally fit way to address potential and current ethical problems. Directly conducting some form of research with people (best from diverse backgrounds) would not only personalize the algorithms behind NLP, but also open up ethical discussions around the technology as a result of the direct involvement of human subjects. Implementation of vernacular would make it more natural and inclusive. The people that would be most affected by these suggestions would generally be people who speak in heavy vernacular or who speak the processed language as a second language.  I believe that is what is most up in the air in the field of NLP.  As with any emerging technology, the developers end up holding a social responsibility, dictating how this technology will be us...