Breaching the Surface
To a person behind the screen at a company coding algorithms, the massive computers filled with user data, and advertisers paying to push their products onto social media platforms, I am nothing more than what information they can gather from me and what I reveal about myself online. This is the information that gets collected, stored, and used to create a computer-based model for me that can predict my behaviors and my personality more accurately than humans can (Youyou et al. 2015). As an avid social media user, it is absolutely certain that my computer-based model is accurate and nuanced to me. But how much do they actually know? With this question in mind, this essay will explore my profile as an Instagram user and how I am viewed by the app, then I will analyze the more hidden parts of who I am–my values, biases, and identity. Finally, I will discuss the intersection of these two spheres and the implications it has.
On Instagram, I have not one, but three accounts. My main account is what I want people to see–my followers include high school friends, college friends, and acquaintances. I have a spam account for my close friends where I “photodump” my silly photos or repost funny posts. And finally, I have an art account where I post my drawings under an online alias. Together, these three online identities contribute to my computer-based model, about who I am to companies, brands, and developers.
In signing up for my Instagram accounts, I provided basic information about myself to the app. My name is Alicia Zhang, I am Chinese, my birthday is February 2nd, 2003. Nineteen years old. Through my interactions on the app, more information can be deduced about me. I’ve lived in Neenah, Wisconsin my whole life, and attended Neenah High School, surrounded by caucasians. Currently, I am a student at UC Berkeley, graduating in 2025. I also follow influencers, my favorite bands and singers, actresses, artists, as well as science and Christian pages. My hobbies include art of all kinds: digital art, painting, embroidery, and music (Appendix Fig. 1). Recently I have been interested in ‘broken humor’ posts, skincare, and jewelry.
The Instagram “explore” tab is how the app targets me as a user to show me content they think I would enjoy. In a video made by Adam Mosseri, the head of Instagram, he explains that the algorithm looks at information about a certain post and your interactions with the poster, your activity, and the content you interact with. The algorithm then guesses how likely you are to interact with the post, and arranges your explore feed appropriately (Adam Mosseri). Across all my accounts, my explore page matches with the type of content I interact with on each account. My main account’s explore page shows various Christianity posts, viral reels from Asian creators, artwork, and a variety of funny posts. On my spam account, I save posts about earrings and skincare and follow humor accounts, so my explore page looks similar. And for my art account: arts and craft content (Appendix Fig. 2). Even the Instagram shop tabs across all accounts are curated to my interests, from jewelry to clothing to sticker art. These examples all demonstrate who I am to an algorithm and as a user of a platform.
While an algorithm is capable of predicting my interests, I am so much more than just my interests. I am Alicia Zhang–an aspiring scientist, among other things. Since I was young I always had a fascination with the complexity of biology. Now that I’m older, I’m especially interested in biomedical research and the impact that I could make on the lives of those who are sick. Within the scope of this class, I want to research the political and socioeconomic impacts of the polio epidemic and the polio vaccine–I also want to learn more about the effects that diseases can have beyond an individual life to learn the most effective way to develop and implement a widespread treatment. Outside the lab and the classroom, I am a fairly mellow person. In a crowd of my favorite people, I’m one of the loudest, but also an introvert and a homebody. It’s difficult to get to know me on a very personal level, for although I am friendly, I am a closed book.
Beneath my composed appearance are the deep thoughts and experiences that shape me to be who I am. I value achievement, belonging, dependability, efficiency, and happiness. Being surrounded by people who accept me and whom I can depend on is what contributes to my happiness the most. And my value for success comes packaged with my enneagram–a three. While my strive for excellence has certainly paid off in terms of receiving recognition, the trade-off is the immense pressure I put on myself (Appendix Fig. 3). Since middle school, I worked hard to achieve good grades and now that I’m in college, that pressure to do well has evolved into a fear of failure. However, I can think clearly and logically about my problems so that I can find a solution or adapt to the issue. Not every decision gets made with clear logic of course, thanks to my cognitive biases that neuropsychiatrist Dilip V. Jeste includes in his article “Wiser”. I relate most to loss aversion and escalation of commitment, both of which play into my value of efficiency. Loss aversion is when you perceive loss as more severe than gains, and escalation commitment is the tendency to stick with a decision because you have already committed time (Jeste 2020). I try to avoid failure as much as I can, since that means I have to start over and try again. Likewise, and sometimes contradictorily, if I have invested time into a failing project, I continue working in an attempt to ‘save’ the project.
Aside from struggles connected to schoolwork or decision-making, I also wrestled with my identity as a Chinese American. According to the US 2020 Census for Wisconsin, Asians make up 6% of the population, while white people is 61.6% (America Counts Staff 2021). Within my county, Winnebago county, the percentage of Asians drops even further to 3.3%, and white people increase to 86.2% (Appendix Fig. 4). Growing up with white peers at school, I felt ashamed to be Chinese. I was embarrassed by my last name, my mom’s heavy Chinese accent, and the ‘exotic’ foods I ate daily. Even though I also grew up attending a Chinese church, I still identified more with my white friends at school–American Chinese, not Chinese American (Appendix Fig. 5). It was not until high school that I finally began to embrace my ethnicity, when all the east Asian students at my high school were invited to a meeting to talk with school staff and share our experiences as minorities. Hearing stories from students who looked like me and sharing my own made me feel heard, and I felt more connected and confident in my ethnicity through our shared experiences. At Berkeley now, where the student body is 30% Asian, I am even more strengthened by my background and culture (Deloittle, and Datawheel).
Something that I discovered as I unravel these two parts of myself–what algorithms see versus what I see–is that targeted products and social media are more geared, at least from my experience, towards my interests, which more or less are not affected by my demographic. My interests change fairly rapidly, and as such, my Instagram explore has shifted to match those interests. Between my phases of obsession with bullet-journaling, video games, nail art, and friendship bracelets, my explore page at each interest looks like it could belong to somebody else. I find that even though Instagram has tracked all my basic information, especially location, race, beliefs–and even my personality perhaps–the only content I see typically just reflects my interests at the time. Aspects of myself that are more nuanced, complex, and hidden are unable to be targeted. The current state of artificial intelligence is not able to detect models or patterns that influence our values and decisions; in order for this to occur, researcher Kaj Sotala from the Center of Long-Term Risk emphasizes that “AIs will need to have models that allow them to roughly figure out our evaluations in totally novel situations, the kinds of value situations where humans might not have any idea in advance that such situations might show up” (Creighton). This further emphasizes the gap between humans and computers. What still sets us apart are the emotions we feel and the things we experience, which a computer cannot relate to and thus cannot incorporate into their algorithm of products. So while Instagram continues recommending me posts about ear piercings and whatever else interests me at the moment, I am content with knowing that the abstract aspects of myself that deeply shape me will not appear on my screen as an Instagram post.
On Instagram, I have not one, but three accounts. My main account is what I want people to see–my followers include high school friends, college friends, and acquaintances. I have a spam account for my close friends where I “photodump” my silly photos or repost funny posts. And finally, I have an art account where I post my drawings under an online alias. Together, these three online identities contribute to my computer-based model, about who I am to companies, brands, and developers.
In signing up for my Instagram accounts, I provided basic information about myself to the app. My name is Alicia Zhang, I am Chinese, my birthday is February 2nd, 2003. Nineteen years old. Through my interactions on the app, more information can be deduced about me. I’ve lived in Neenah, Wisconsin my whole life, and attended Neenah High School, surrounded by caucasians. Currently, I am a student at UC Berkeley, graduating in 2025. I also follow influencers, my favorite bands and singers, actresses, artists, as well as science and Christian pages. My hobbies include art of all kinds: digital art, painting, embroidery, and music (Appendix Fig. 1). Recently I have been interested in ‘broken humor’ posts, skincare, and jewelry.
The Instagram “explore” tab is how the app targets me as a user to show me content they think I would enjoy. In a video made by Adam Mosseri, the head of Instagram, he explains that the algorithm looks at information about a certain post and your interactions with the poster, your activity, and the content you interact with. The algorithm then guesses how likely you are to interact with the post, and arranges your explore feed appropriately (Adam Mosseri). Across all my accounts, my explore page matches with the type of content I interact with on each account. My main account’s explore page shows various Christianity posts, viral reels from Asian creators, artwork, and a variety of funny posts. On my spam account, I save posts about earrings and skincare and follow humor accounts, so my explore page looks similar. And for my art account: arts and craft content (Appendix Fig. 2). Even the Instagram shop tabs across all accounts are curated to my interests, from jewelry to clothing to sticker art. These examples all demonstrate who I am to an algorithm and as a user of a platform.
While an algorithm is capable of predicting my interests, I am so much more than just my interests. I am Alicia Zhang–an aspiring scientist, among other things. Since I was young I always had a fascination with the complexity of biology. Now that I’m older, I’m especially interested in biomedical research and the impact that I could make on the lives of those who are sick. Within the scope of this class, I want to research the political and socioeconomic impacts of the polio epidemic and the polio vaccine–I also want to learn more about the effects that diseases can have beyond an individual life to learn the most effective way to develop and implement a widespread treatment. Outside the lab and the classroom, I am a fairly mellow person. In a crowd of my favorite people, I’m one of the loudest, but also an introvert and a homebody. It’s difficult to get to know me on a very personal level, for although I am friendly, I am a closed book.
Beneath my composed appearance are the deep thoughts and experiences that shape me to be who I am. I value achievement, belonging, dependability, efficiency, and happiness. Being surrounded by people who accept me and whom I can depend on is what contributes to my happiness the most. And my value for success comes packaged with my enneagram–a three. While my strive for excellence has certainly paid off in terms of receiving recognition, the trade-off is the immense pressure I put on myself (Appendix Fig. 3). Since middle school, I worked hard to achieve good grades and now that I’m in college, that pressure to do well has evolved into a fear of failure. However, I can think clearly and logically about my problems so that I can find a solution or adapt to the issue. Not every decision gets made with clear logic of course, thanks to my cognitive biases that neuropsychiatrist Dilip V. Jeste includes in his article “Wiser”. I relate most to loss aversion and escalation of commitment, both of which play into my value of efficiency. Loss aversion is when you perceive loss as more severe than gains, and escalation commitment is the tendency to stick with a decision because you have already committed time (Jeste 2020). I try to avoid failure as much as I can, since that means I have to start over and try again. Likewise, and sometimes contradictorily, if I have invested time into a failing project, I continue working in an attempt to ‘save’ the project.
Aside from struggles connected to schoolwork or decision-making, I also wrestled with my identity as a Chinese American. According to the US 2020 Census for Wisconsin, Asians make up 6% of the population, while white people is 61.6% (America Counts Staff 2021). Within my county, Winnebago county, the percentage of Asians drops even further to 3.3%, and white people increase to 86.2% (Appendix Fig. 4). Growing up with white peers at school, I felt ashamed to be Chinese. I was embarrassed by my last name, my mom’s heavy Chinese accent, and the ‘exotic’ foods I ate daily. Even though I also grew up attending a Chinese church, I still identified more with my white friends at school–American Chinese, not Chinese American (Appendix Fig. 5). It was not until high school that I finally began to embrace my ethnicity, when all the east Asian students at my high school were invited to a meeting to talk with school staff and share our experiences as minorities. Hearing stories from students who looked like me and sharing my own made me feel heard, and I felt more connected and confident in my ethnicity through our shared experiences. At Berkeley now, where the student body is 30% Asian, I am even more strengthened by my background and culture (Deloittle, and Datawheel).
Something that I discovered as I unravel these two parts of myself–what algorithms see versus what I see–is that targeted products and social media are more geared, at least from my experience, towards my interests, which more or less are not affected by my demographic. My interests change fairly rapidly, and as such, my Instagram explore has shifted to match those interests. Between my phases of obsession with bullet-journaling, video games, nail art, and friendship bracelets, my explore page at each interest looks like it could belong to somebody else. I find that even though Instagram has tracked all my basic information, especially location, race, beliefs–and even my personality perhaps–the only content I see typically just reflects my interests at the time. Aspects of myself that are more nuanced, complex, and hidden are unable to be targeted. The current state of artificial intelligence is not able to detect models or patterns that influence our values and decisions; in order for this to occur, researcher Kaj Sotala from the Center of Long-Term Risk emphasizes that “AIs will need to have models that allow them to roughly figure out our evaluations in totally novel situations, the kinds of value situations where humans might not have any idea in advance that such situations might show up” (Creighton). This further emphasizes the gap between humans and computers. What still sets us apart are the emotions we feel and the things we experience, which a computer cannot relate to and thus cannot incorporate into their algorithm of products. So while Instagram continues recommending me posts about ear piercings and whatever else interests me at the moment, I am content with knowing that the abstract aspects of myself that deeply shape me will not appear on my screen as an Instagram post.
Appendix
Works Cited
Adam Mosseri [@mosseri]. How the “Algorithm” Works. Instagram, 23 Jun. 2021, https://www.instagram.com/p/CQdxvdNJ_sC/?utm_source=ig_embed&utm_campaign=embed_video_watch_again
America Counts Staff. "Wisconsin Population Increased 3.6% Since 2010." https://www.census.gov/library/, 25 Aug. 2021, www.census.gov/library/stories/state-by-state/wisconsin-population-change-between-census-decade.html#:~:text=Race%20and%20ethnicity%20(White%20alone,or%20More%20Races%2010.2%25).
Creighton, Jolene. "How Can AI Systems Understand Human Values?" Future of Life Institute, 14 Aug. 2019, futureoflife.org/2019/08/14/how-can-ai-systems-understand-human-values/?cn-reloaded=1.
Deloittle, and Datawheel. "University of California-Berkeley." datausa.io, datausa.io/profile/university/university-of-california-berkeley#:~:text=Enrollment%20by%20Race%20%26%20Ethnicity&text=The%20enrolled%20student%20population%20at%20University%20of%20California%2DBerkeley%20is,Hawaiian%20or%20Other%20Pacific%20Islanders.
Jeste, [Excerpts on] Cognitive Biases in Wiser, 2020
Neenah Joint School District. "Back to School." Dialogue, [Neenah], 2020.
Youyou, Wu, and Michal Kosinski. "Computer-based personality judgments are more accurate than those made by humans." Psychological and Cognitive Sciences, vol. 112, no. 4, 12 Jan. 2015, PNAS. www.pnas.org/doi/full/10.1073/pnas.1418680112.
Zhang, Alicia. “Instagram Explore Page.” Screenshot, 20 Sept. 2022.
Zhang, Alicia. Various photographs.
America Counts Staff. "Wisconsin Population Increased 3.6% Since 2010." https://www.census.gov/library/, 25 Aug. 2021, www.census.gov/library/stories/state-by-state/wisconsin-population-change-between-census-decade.html#:~:text=Race%20and%20ethnicity%20(White%20alone,or%20More%20Races%2010.2%25).
Creighton, Jolene. "How Can AI Systems Understand Human Values?" Future of Life Institute, 14 Aug. 2019, futureoflife.org/2019/08/14/how-can-ai-systems-understand-human-values/?cn-reloaded=1.
Deloittle, and Datawheel. "University of California-Berkeley." datausa.io, datausa.io/profile/university/university-of-california-berkeley#:~:text=Enrollment%20by%20Race%20%26%20Ethnicity&text=The%20enrolled%20student%20population%20at%20University%20of%20California%2DBerkeley%20is,Hawaiian%20or%20Other%20Pacific%20Islanders.
Jeste, [Excerpts on] Cognitive Biases in Wiser, 2020
Neenah Joint School District. "Back to School." Dialogue, [Neenah], 2020.
Youyou, Wu, and Michal Kosinski. "Computer-based personality judgments are more accurate than those made by humans." Psychological and Cognitive Sciences, vol. 112, no. 4, 12 Jan. 2015, PNAS. www.pnas.org/doi/full/10.1073/pnas.1418680112.
Zhang, Alicia. “Instagram Explore Page.” Screenshot, 20 Sept. 2022.
Zhang, Alicia. Various photographs.