An interesting source I found from Google talking about how YouTube's recommendations work: https://support.google.com/youtube/answer/16089387

Youtube apparently recommends videos based on a combination of their interests and behavior. The algorithm considers factors like watch history, search history, likes, dislikes, and watch time. All of these work together to predict which videos someone might enjoy

One interesting detail I learned was that YouTube doesn't just measure how long someone watches a video, but also uses satisfaction surveys to determine whether users enjoyed the content too, which helps with distinguishing high performing videos from low performing videos. I previously thought that YouTube gathers all their recommendation info from how people watch the videos themselves, opposed to manual surveys.

I've had experiences with instagram and youtube recommending content that matched my interests, even when i hadn't searched for them. A good example of this is about AI and other technologies. After watching a few videos about AI and technology, my fyp started giving me new research, tech products, and related content that I found interesting. One thing that was also interesting was that I noticed recommendations improved over time as I interacted with more posts.

https://www.nytimes.com/interactive/2022/02/18/arts/alt-text-images-descriptions.html

Summary: This is an article from the NYTimes that talks about how the internet could be made more accessible to people who are blind or visually impaired using alt text, which are written descriptions of images that screen readers can interpret.

One Interesting detail I learned was that in February of 2021, only around 40% of homepages included alt text. I found it a little surprising because it's a relatively simple accessibility feature, but much of the internet isn't designed with it in mind.

With social media, disability can be viewed in terms of how acessible to different users. Social media platforms are often designed with certain users in mind, which can unintentionally exclude people with disabilities. For example, videos without captions can exclude people who are deaf or hard of hearing, while platforms that don't support screen readers can make it difficult for those who are visually impaired. From an ethical perspective, I believe that platforms have a responsibility to design around different disabilities rather than expecting users to adapt to systems that weren't built with them in mind.

An additional source I found: https://www.pewresearch.org/internet/2021/01/13/the-state-of-online-harassment/

This interesting article from Pew Research talks about different forms of harassment, name-calling, embarrassment, stalking, threats, etc. It also talks about how online harassment has changed over time, finding that more severe forms of harassment have become increasingly common, and how platforms are trying to address it.

One detail I found interesting is that about 41% of American Adults reported experiencing some form of online harassment, which shows the extent to how much it spreads. Additionally, the research found that severe forms of harassment have increased even though the overall percentage of people experiencing harassment has remained relatively stable, which makes me wonder how out of control these platforms are in order to observe behavior like this.

I feel like the best way to respond to trolling is to just ignore it. Oftentimes, most trolling is merely just bait in order to provoke a reaction from people. If fewer people give trolls the reaction that they want, they'll probably die down. The best move is to just report it to the platform you're using and then just move on with your day. A good example of where this worked is the comedian that was using twitter to impersonate the FBI by saying the assassination of MLK was done by the FBI when it wasn't. Twitter had suspended his account, which likely happened because he was reported for doing so.

An interesting article I found on the death of authenticity in the age of AI and the internet: https://medium.com/@patrickryan333/the-death-of-authenticity-how-social-media-turned-us-into-brands-5a33af22240e

This article talks about how social media has changed the way people present themselves. Often the algorithms powering these platforms encourage people to create carefully curated online identities rather than showing their authentic personalities. People are increasingly treating themselves like brands. They're selectively positing their happy moments, good looking photos, accomplishments, etc. while generally hiding their flaws and struggles. This tends to create an unrealistic expectation of what people's lives actually look like.

One detail I found interesting from the article is just how much social media has changed people's perception of the platforms themselves. People have now become focused on curation of experiences rather than purely just sharing experiences because of how algorithms dictate what types of content is shown to people.

In my personal experience, talking to people over the internet is very different from talking to people in real life. Generally speaking, people on the internet don't know much about you and often can't see your authentic self. I think this creates a disconnect that allows people to feel more confident interacting with others online than they would with those same people in person. Social media also gives you time to think about your responses or comments, as opposed to real-life conversations where you often have to respond on the spot. Personally, I don't think this necessarily compromises authenticity, as people may simply feel more comfortable expressing parts of their personality online that they would otherwise hesitate to show in person.

I'm comparing achievement/event sharing between Instagram and Twitter/X.

Instagram has image focused design, which makes it easier for users to visually showcase their achievements. The app is centered around sharing photos and videos, so users can provide direct visual evidence of their accomplishments, which allows them to be more engaging and immediately recognizable.

Twitter meanwhile, is very text-focused which makes it easier for users to describe and elaborate on their achievements. While users can attach images, the platform's emphasis on text and threaded posts allow them to provide things like the process behind the accomplishments, and connect multiple achievements into a single narrative. This makes Twitter useful for achievements that require an additional explanation.

Source - https://medium.com/@sophiebonheim/the-anatomy-of-a-good-social-media-post-62074d5fe6d6

Summary: The Medium article discusses how principles of UX and visual design can be applied to social media posts to improve engagement. I took note that the article contains details that are often used in a post that's maximized for engagement such as attention-grabbing hooks, hashtags, spacing, calls to action, and visual hierarchy. I learned that some posts are more optimized for engagement from others and this source explains how these design choices can make content easier to read, more accessible, and more engaging for users.

Detail: One detail I found really interesting was how even something as simple as the formatting and placement of hashtags can influence how people interact with a post. I find this relevant to the class and the topic related to data because it shows how information presented in a specific way can influence user behavior. Even seemingly small design choices can affect what content gets attention and how people engage with it.

I think this is interesting to consider how social media platforms evaluate the success of their algorithms. In my own research, I found platforms often prioritize metrics like engagement and watch time. But this often comes at the risk of often overlooking negative effects on users' mental health or wellbeing. This seems like a form of pernicious ignorance because we're defining success based on the data that's easiest or most beneficial to measure, rather than considering the full consequences. It makes me wonder how much of this ignorance is intentional versus simply a result of the way we design our systems.

Another interesting source I found on the Dead Internet Theory from Harvard: https://ui.adsabs.harvard.edu/abs/2025arXiv250200007M/abstract

Summary: The Dead Internet Theory is the idea that social media is becoming dominated by bots, AI-generated content, and corporate interests. This study looks at where the theory came from, its main ideas, and how it affects social media. It focuses on how platforms prioritize making money and keeping users engaged over real human interaction, making the internet feel less authentic and trustworthy.

Detail: One interesting detail from the paper is that an estimated 40–60% of web traffic comes from bots. I also learned the extent to which bots on these websites also artificially increase likes, comments, and shares, making online communities appear more active than they actually are. While I already suspected this on platforms like X. I feel like this study brought to life just how much of a problem it truly was.

People's expectations for a bot are different from those for a normal user, given that it's an automated response rather than one from an actual human. If someone knows that content comes from an actual human, it adds to its trustworthiness. On the flip side, a bot indicates that someone is trying to gather, scrape, or automate the delivery of information, which could make people trust it a lot less.

I find it really interesting how some ancient ethical frameworks are still heavily used and seen today, especially in different countries. While Confucianism was primarily adopted in China and Eastern Asia, I've noticed a very similar type of culture in how my parents grew up in India. Sometimes, traditions and certain frameworks are so strong that they're not only seen in multiple cultures, but even make it into the modern day.

Curiously, I've noticed a difference in Western culture's take on this versus that of many Eastern countries. While respect for elders and parents is still really important in Western culture, it seemingly isn't emphasized as strongly or traditionally as it is in Eastern cultures even today.

I believe that tech workers should always evaluate the ethical considerations of what they're building throughout the entire process, from conceptualization to development to A/B testing. This is because, oftentimes, we don't realize the destructive potential of a specific technology until it's too late.

I think the responsibility of tech workers is to consider and address these potential issues very early in development so that we can fix them before they become serious problems.