# Reading Responses (Set 2)

#### Table of Contents
1. [Online Dating](##Online-Dating)
2. [Advertising](##Advertising)
3. [Five Stars](##Five-Stars)
4. [Artificial Intelligence](##Artificial-Intelligence)
5. [Algorithmic Bias](##Algorithmic-Bias)

## Online Dating

Without Tinder, how did anyone ever meet each other? Relationships began through set ups, friends and family members held cupids arrow and played the game of matchmaker. This was how relationships began for thousands of years, until technology revokes cupids power from the people and took it for itself. Dating apps became popular in the early 2000s, and by 2018, [10% of the couples featured in the New York Times Wedding Announcements met on dating apps](https://www.theatlantic.com/family/archive/2018/12/tinder-changed-dating/578698/). These artificial matchmakers allow people to meet mates that they probably would have not met without the assistance of the web. Marginalized communities, such as the LGBTQ+, benefited immensely from the introduction of online dating, because they often were not surrounded by viable dating options. Online dating allows these communities to find like-minded people outside of their communities. Even just in general, dating apps allow for expansion outside of inner circles and provides opportunities to integrate new, strong ties. But on the other hand, there is now an expectation that you will date outside of your social circle, isolating dating from social life. In the past, people would go out to bars with the intention of meeting someone, but now people go out with friends just to have fun. Hoping to meet a future spouse stays contained on your phone.

Dating apps also provide a breeding ground for trolls and dangerous situations. Just like on any online forum, there is no way to know whether or not the person on the other side of the screen is real. Furthermore, with dating apps, if you plan to meet up with someone through an app, there is no 100% foolproof way to know that the person you are meeting up with truly is who you think they are and that they have pure intentions. Dating apps certainly have their pros and cons, but they have integrated themselves into today's society and most likely will not be going anywhere anytime soon.

## Advertising
The route from an advertisement to the purchase of a good or service is now faster than ever, thanks to the internet. Before the internet, advertisements were meant to stick in consumers heads and influence later purchases made within stores. People repeatedly saw billboards and heard jingles, and were influenced to buy things in stores. Now, pesky pop up ads and banners allow impulsive purchases to be one click away, linking consumers to any product imaginable. The internet also broadened the scope of who can advertise.

Classic ways of advertising online include banners, popups, and floating advertisements, all things that most people who have accessed the internet have seen. Over the past few days, I have paid attention to the most prominent types of advertisements, and I would say that I see mostly banner ads, map adverts found searching on my maps app for directions, and influencer sponsorships.

Influencer sponsorships are the main way that micro-influencers on social media make their money, earning hundreds, if not thousands of dollars for a singular post about a product. Brands reach out to these people with followings and "gift" them their products in exchange for a post. Although this method does not provide the straightforward targeting and tracking capabilities that a typical advertisement embedded on a website would produce, it is one of the most effective ways of advertising to the younger generations. Influencers got their title of "influencer" for a reason - they influence the behaviors of their fans and followers. Brands take advantage of this by using these micro-celebrities as a means to spread awareness of their product.

One of the most fascinating parts of the reading to me, however, was the mention of "advertisement fatigue," as people see so many ads on a daily basis that they are likely to ignore advertisements. This concept can already be seen in video media, for example people fast forward commercials on TV or click "skip ad" when watching YouTube. With advertising fatigue, it will be interesting to see how advertising as a concept changes in the future.

## Five Stars!

Information asymmetry states that one participant in a transaction tends to know more about what is happening than the other. As ordering online gained ease, the chances that the product you ordered online might be different from the product that shows up at your door. Who is holding the sellers accountable? Customer reviews provide knowledge to the consumer about what they plan to purchase. However, there is no way to know truly who is on the other side of the screen writing a review. Many times, it is really a consumer who really enjoyed a product. On the other hand, it can be the owner of a product, desiring to boost their product through positive reviews, or maybe even a troll who intends on bothering other people. Either way, reviews must be taken with a grain of salt, on the off chance they are fake, as "researchers estimate that between 10 to 30 percent of online reviews are fake" (Reagle, 2019).

The emergence of TikTok shop brought a new clan of reviewers to the internet. TikTok "gifts" certain amounts of money to creators with followings to spend on their in-app shop. The catch is that the creators must make a video with the product they choose to purchase. Furthermore, when creators link these products to their videos, they make a small cut of the sales. At the bottom of TikTok videos where products are linked, there is an "eligible for commission" banner, but still, it is unclear to most users on the app what this banner means. Some creators stay transparent and inform their followers that they received these products for free and make commission off the link, while others stay silent and leave their viewers in ignorance. Some creators even use this feature as the main focus of their account, basing their content on reviews of TikTok shop products. Regardless, there is no way to know if these reviews are genuine, or if the creators want to keep making money.

## Artificial Intelligence

One of my current favorite ways to find new music is through utilizing chatGPT's knowledge base. I provide it with a prompt including a song I like or a vibe I'd like to hear in my music, and it spits out a good list of recommendations based on my input. My feedback on these responses actually teaches the AI to improve its answers for other people.

ChatGPT and other artificial intelligence aim to reproduce "similar but distinct" content based on work humans have already created. Users are free to use the sites, not because the companies are kind, but because user interaction with the AI increases its performance. Learning from human input has its disadvantages, though, leading to AI feeding into preexisting human created stereotypes that are prominent online.

AI has its advantages, assisting in email drafting, ideation, and content creation. However, there are negatives as well, especially when it comes to generation of harmful content. The AI produces text and images via the database it has access to. The larger the scope of data available to the AI, the more likely mistakes are bound to happen, causing the AI to go haywire. An example of this is the distinction between the Bing AI, Sydney, compared to the Spotify DJ AI. The Spotify DJ AI hardly malfunctions because it's data access is limited to music. On the other hand, the Bing AI had access to most of the data on the internet, meaning it was able to read information online about itself. This ability to self-read drove the AI to have alter-egos, confess it's love for users, and even tell users that they have been bad.

## Algorithmic Bias

"Models are opinions embedded in mathematics" (O'Neil, 2016, p. 21). Models and algorithms are often regarded as unbiased forms of decision making tools. People use models for just about everything these days, from determining personality types all the way to ranking the best cities to raise kids in. While users may assume that these models serve simply to tell them the truth, they forget that there are people and contexts behind the algorithms, leading the answers to one outcome or another.

Models plagued the prisons systems, as self-identification quizzes became popular to determine sentencing length to remove race as a determining factor. Those on trial would fill out a survey that would then be translated into an appropriate sentence length based on their answers. However, some questions asked within the survey are those that would not be allowed in court, since they feature no relevance to the current case. For example, participants would be asked about their family's criminal history. While in trial, this information is irrelevant. Asking for it on the survey, on the other hand, insinuates that these prospective inmates are linked to the actions of their families. It also enforces existing stereotypes. Less wealthy areas with more crime are more likely to have police roaming around, meaning that citizens are more likely to have run ins with the cops, whether or not they committed an illegal act.

Algorithmic bias looms on the internet as well. A few years ago, users found that searching for "professional hair" yielded images of white people, while "unprofessional hair" showed images of black people. While these racial biases have since been corrected, they show that biases found in the human world transfer over to the internet.

Looking at the community creating most of the algorithms and AI, some of the reasons for these biases become clear. The creators of these technologies are predominantly white men. Within this space, without equal inclusion of everyone, biases will continue to exist.