IP#3 - Algorithms of Everyday Life

1. Informed Consent: Historically, informed consent referred to the clear and voluntary indication of preference or choice, typically after being apprised of the risks and benefits. This could mean signing a waiver to understand the risks of skydiving before doing so. In the digital age, with algorithms playing a pivotal role in data collection and processing, the concept of informed consent becomes more complex. Users often provide consent without fully understanding the depth and breadth of data collection, and how algorithms might use this data. Neyland's work underscores the intricacies of these interactions, emphasizing the need for transparency and clarity. For example, many people don’t fully read end user license agreements before they click ‘I agree’. In another example, a surveillance system might be using our facial data to track where we are at any given time, even though we may not have been informed properly that it is doing so. The idea is that algorithms do affect human behaviors and, in some cases, dictate human behavior. If that is the case, and it is, then it becomes increasingly important that we understand how the algorithm makes its decisions before we subject ourselves to the whims of the algorithm.

1. Fair Use: While the term "fair use" traditionally pertains to the right to use copyrighted material without permission for specific purposes, in the algorithmic context, it can be extended to consider how data is used. Algorithms can repurpose data in ways that might not align with original intentions, leading to ethical dilemmas. A recent example of this is how the Midjourney AI just scrapes all images from the internet and uses these images to generate ‘original’ ones. Traditionally, fair use only applies if the copyrighted material is being used in a remix (commentary, satire, non-commercial uses etc.) but algorithms have paved the way for any and all copyrighted material to be used as a data input that trains the algorithm. This puts the laws of fair use to the test. AI is exceptionally good at paraphrasing, or not copying the syntax, but instead copying the idea. It is clear in all academia that we should cite our sources if we are using another person’s ideas. However, how do we do that with art? How do we do that when the AI doesn’t even know where it is pulling the idea from? How do we enforce copyright law? All these call for a redefinition of fair use, either making it more restrictive, at the risk of stifling innovation and creativity, or relaxing the requirements for fair use, at the risk of rampant plagiarism and the commoditization of art.

1. Discrimination and Net Neutrality: Discrimination in the digital realm can be amplified by algorithms that inadvertently reinforce biases present in the data they're trained on. Net neutrality, the principle that all internet traffic should be treated equally, can be compromised by algorithms that prioritize certain types of content over others. Neyland touches on the implications of these algorithmic decisions and their broader societal impact. For example, Google and YouTube isn’t incentivized to show the most accurate and reliable information, the platforms are incentivized for session times, watch times and retention because that’s how they can show user more ads, and in turn, make more money. Such is the case for every platform we use that we are not paying for. It used to be that the news was supposed to present an unbiased account of important events relevant to us. Nowadays, the news is about which headline can get the most clicks. While an algorithm doesn’t discriminate outrightly, its ability to shape our preferences by controlling our content diet and lead humans down certain inclinations is downright dystopian, and something we might not even be conscious about.

1. Personalization: Once a term that implied a tailored experience based on individual preferences, in the algorithmic world, personalization often means a hyper-curated experience based on data-driven insights. While this can enhance user experience, it also raises concerns about privacy and the potential for creating echo chambers, where individuals are only exposed to content that aligns with their existing beliefs. Target figured out how to predict pregnancy in women as early as 2012 so they can send targeted flyers to these women, as mentioned in Charles Duhigg’s book The Power of Habit. While personalization is good if it gives us what we want according to our preferences, algorithms have gotten increasingly competent at shaping our preferences and behavior. The astounding predictive ability of algorithms that do so to profit in the name of personalization implies that the word personalization should be redefined. Is it still personalization if an algorithm caters to our preferences before we formally tell the algorithm our preferences? Whether our user experience is actually being enhanced, or the algorithm is grooming us to prefer a certain user experience would be an interesting topic to research.

1. Friend: In the pre-digital age, a friend was someone with whom you shared a personal relationship. Today, in the era of social media algorithms, the term has been commodified. Algorithms determine which "friends" appear on your feed, often based on engagement metrics rather than personal closeness. Neyland's exploration highlights how such algorithmic decisions can shape our perceptions and interactions. Facebook and Instagram don’t show us our feeds based on the quality of our personal relationships; they do so based on their understanding of what would keep us on their platform. However, doing so has also shaped our understanding of who our friends are. If we maintain relationships by maintaining a two-way connection, then who the algorithm shows us will ultimately affect who we keep in touch with, and as a result, who is a friend. With the coming of ChatGPT, and Elevenlabs, we might even become ‘friends’ with an AI in the future. Our future inability to tell AI apart from human would have serious sociological implications and would call for a renewed understanding of what having a friend actually means.

References

Neyland, D., Springer Social Sciences eBooks 2019 English/International, OAPEN, DOAB: Directory of Open Access Books, SpringerLink (Online service), & SpringerLink Fully Open Access Books. (2019;2018;). The everyday life of an algorithm (1st 2019. ed.). Springer International Publishing. https://doi.org/10.1007/978-3-030-00578-8