Thank You!

Thank you to everyone who has read, shared, and engaged with The AI Chronicles this year. This final edition of the school year highlights four student final papers from our AI course. Your readership has helped make this work feel meaningful beyond the classroom, and I’m grateful for it.

Table of Contents:

Lily Tym, "In Defense of Friction: Preserving Human Growth in the Age of AI"

Natalie Wiesner, "Minds, Machines and Medicine: How Artificial Intelligence is Reshaping Surgery, Diagnostic, and the Future of Healthcare"

Noah Ivanchikova-Crouse, "Balancing Innovation with Regulation"

Angelina Castillo, "The Ethics of Artificial Intelligence: How AI is Reshaping Human Life"

Lily Tym, "In Defense of Friction: Preserving Human Growth in the Age of AI"

Introduction

The use of Artificial Intelligence, specifically LLMs, has transformed from something used by only a percentage of the population into a feature of nearly every business, website, and even personal life. It has evolved into something that is impossible to ignore. Though there may still be some individuals who are able and choose to avoid AI in their own personal use, the odds that they have unwillingly come across it in some format are extremely high. An even more surprising aspect than just the widespread usage of AI, to me at least, is just how fast this shift has occurred. OpenAI first released their original model, GPT-1, to the public in 2018. It hasn’t even been a decade, yet this technology is already a global sensation and pretty much inescapable. I wouldn’t necessarily claim that the problem is the innovation itself—in fact, I think that piece of the puzzle is actually inevitable given human curiosity and evolution—but there are harmful consequences that come with the overuse of AI and we must consider them as we move into the future.

A major argument that has been used time and time again is that the overuse of Artificial Intelligence harms critical thinking skills. While this is true—and backed up by research—I think this claim fails to address the more detrimental effects that AI has the potential to have on humanity as a whole. Since it does affect the way we think, it therefore has the capacity to shape the way we see reality. This then has effects on society through the way our policies are shaped, the way in which we express our ideas through research and writing, and last but certainly not least, our education system—something that has been, for generations, promised to be the root of societal improvement. These alterations are happening right in front of our eyes, and there is no way of confirming when these developments will ever slow down, or if that’s even a possibility. Countries and groups with immense power, specifically in the U.S. and China, continually advocate their support for the growth and development of these new technologies, and it has become essential for their success. Many companies have made it clear that it is a new priority of theirs to implement AI in a way that is efficient and productive for their own interests, and as of right now, there are no real claims backed with an action plan against the usage of AI. 

If there are so many plausible possible drawbacks of this AI epidemic, why are we allowing this to happen? Personally, I think it is because the easiest way to deal with these changes is to passively witness them rather than acknowledging the detrimental impacts. The way I see it, no one person can logically be held accountable for the effects AI has had and will continue to have on society. To be fair, there are specific founders and developers that created Artificial Intelligence in the first place, but as a society we have continually embraced these innovations with open arms. For generations, technological innovation has been highly prioritized, and it is therefore unfair to pinpoint the blame to just a concentrated group. Even though there is a minority of individuals who are reaping most of the benefits, the system has indirectly encouraged this process for years. That being said, I think the solution also lies in the hands of society. I say this in relation to one of the largest impacts of AI, which has reshaped our daily lives.Human existence no longer requires the hardships or “friction” that it once did and as a result, it is influencing everyone. The importance of friction for human development and growth needs to be recognized on not just an individual level, but as a widespread philosophical change. Perhaps a possible remedy that could genuinely make an impact is to slowly but consistently implement more friction into our lives. Maybe then we can realize that an overuse of AI is harming not only our intellectual capacity as a species, but also our morality, and maybe then we can begin to make the active decision to reject this practice.

The Emergence of AI-assisted Writing

When I entered high school as a freshman, the possibility of AI being used in the sphere of writing was already a hefty possibility. I distinctly remember forwarding a video on social media that indirectly asserted my superiority over other students for refusing to use AI in my coursework. My perspective surrounding AI has definitely shifted since that period in my life, and I have since then come to terms with the fact that AI is a useful tool in learning. For the purposes of my own education, I currently utilize it for more formulaic tasks such as building study plans and creating practice problems. On some level though, I still stand by the fact that AI wasn’t meant to be used for writing, and it is fundamentally harmful to the students who do use it for this purpose. While it is true that AI has the ability to assist with grammatical errors and quick fixes, it can harm critical thinking skills when it is overused, which are crucial to the writing process itself. In this way, I believe that the growing lack of critical thinking is the basis of our current lack of friction in society. 

Beyond critical thinking, AI writing is just simply not as meaningful as human writing, and it never will be. LLMs specifically are trained on data with certain biases present throughout the process, inherently shrinking the amount of information they have access to. Compared to human writing, AI lacks the sense of real life experiences. The process of human writing contains a certain amount of meaning behind it that AI will simply never be able to replicate, because it hasn’t lived a life of experience with a multitude of perspectives like humans have. Furthermore, it takes away from the value of what real, human struggle feels like when you’re part of a larger society. An Artificial Intelligence model has no hardships to overcome. In other words, LLMs are not able to experience friction in their ‘lives’—if you can even call it that—which is another reason why it lacks meaning. Of course, it is up to the reader to decide whether a piece of writing has “meaning” behind it, but Eve Fairbanks makes what I believe to be an excellent argument regarding ‘meaningful writing’: she claims that the creation and process behind the writing is what truly has the ability to give it meaning. To support this, she reminds readers that “when human beings write, we judge ourselves; we stop; we backtrack” (The Atlantic). Most LLMs do not backtrack on their work, and the sheer speed of the technology is enough to prove this. As a result, this part of the process is lost when using AI to write or even edit your writing. As someone who enjoys writing, I can confidently confirm that the actions of drafting, revising, and editing is particularly important to the satisfaction I feel when I finish an essay or even a personal memoir. Using AI in the writing process isn’t necessarily morally incorrect to do—and there are definitely plausible arguments to why it is sometimes neccessary—but it is no way helpful to developing real writing skills or, more importantly, critical thinking. 

Even if you happen to be one of the individuals who continually refuses to use AI as even a “writing tool,” there is still a clear relationship between the way that an LLM produces written works and the way in which humans produce them. The abbreviation “LLM” quite literally stands for “Large Language Model,” so it is inherent that the written language of AI is a core factor to the general system. Because written language plays such an important role in the work of AI, it has the ability to lead to yet another chain reaction with humans as the leading test subjects. The fact is, even if you do not use AI in your own writing, others do and they will most likely continue to partake in this practice despite the growing societal movement targeted against it. Considering that as the first step in the chain reaction, the second step would be regarding the human reaction to it. People are highly influenced by the media they consume, and if you’re consistently reading AI-produced work, it will eventually begin to influence your writing style as well—even if you are not conscious of this fact. This pattern is being increasingly noticed by linguists, specifically Max Planck at the Institute in Germany, who found that “YouTubers began using words favored by AI chatbots up to 51% more frequently after the release of ChatGPT” (Fortune). Logically speaking, there are two possible causes for this increase. The first possible cause is that the Youtubers themselves are beginning to use AI as an assistive writing tool to write their scripts. Without a doubt this has already begun to happen, but the scale of this metric is still widely unknown. The second, more widespread cause—specifically for the creators who continue to produce honest work in spite of AI—is probably that these creators are being influenced by others in the same space as them who do choose to use AI as a “writing tool.” Logically, there is no way to get confirmation on if these are the exact causes, and it’s almost an impossible theory to test due to the dishonesty and shame many creators may feel from using AI in their pursuit of content creation. Whatever the reason, it is impacting nearly everyone and the impact of LLM wording is becoming harder and harder to isolate yourself from.

Even from personal experience, I know that this chain reaction is real. Every day, I see more and more of the same formulaic writing infiltrating my feed, even from creators I believe refuse the use of AI in their writing space. As a writer myself, I sometimes find myself questioning if my writing sounds like Chat-GPT simply because of the amount of AI-created writing I have willingly and unwillingly consumed. As many researchers have come to realize, it is easy to identify whether something has been written by an AI model. Over and over again, research has confirmed that Chat-GPT has a preference towards certain words and sentence structures. Due to the training data of specific LLMs, they recognize patterns in language that indicate what “appear in texts that have been marked as well-formed, high-quality prose” (NYT). Sam Kriss uses this theory to explain why AI favors structures such as “it’s not X, it’s Y” and their overuse of em dashes in AI writing. While this is useful for the process of determining if something was AI-written or not, this homogenous writing style will continue to exist and perhaps even evolve in the years to come. This homogeneity will also have significant effects on the already increasing lack of friction in our society. Without differentiating editorial point of views to dispute with, there will ultimately be less discourse in the sphere of writing, specifically in academia and journalism. Additionally, in the long-term, this process of confirming whether or not something is human written is only going to become more difficult. I make this claim simply because as more humans are affected by the way AI writes, the closer these LLMs will sound to real human writing. It is simply unrealistic to deny the mutual relationship between human writing and AI writing, and it’s not going away anytime soon. 

The Integration of AI in Education and Research

Artificial Intelligence has heavily infiltrated not only the world of writing, but also the sphere of academic intellect. More specifically, AI’s impact on education and research communities is especially prominent, and perhaps more concerning than its effects on writing. Although writing is a huge part of research as well as education, there are other steps in the general process of learning that are slowly declining in effectiveness due to the increasing use of AI. I say this because while writing is an important element in human thought and a priority in many education spaces, I think that reading and writing has become more optional over time. Nowadays, students are likely to find a summary of an assigned text or ask AI to analyze rather than do the work themselves. In this way, it is largely up to individual choice and partially the teachers to bear the responsibility of providing a proper education that benefits the formation of critical thinking. Again, it is common for students—as well as teachers—to use AI as a “tool” in their research, but is it truly helpful? This comes in many forms, such as brainstorming, finding sources, or even summarizing the information for the assigned research itself. Many researchers have evaluated the effects of students using AI in research specifically, and come to the conclusion that cognitive offloading is one of the major downsides. Cognitive offloading is defined as the process in which physical action is used to reduce the cognitive processes our brain is required to execute in order to complete a task. In simpler words, it is an easier way of “learning” information without deeply understanding it. A less modern example of this is writing information down when studying for a test: it helps with memorization, but it doesn’t encourage the same layered thinking that is required to master a specific concept. When it comes to the use of AI for cognitive offloading, it has been proven to help “simplify the process of finding answers,” but it doesn’t “encourage the depth of engagement that comes with searching through a range of diverse sources and critically evaluating them” (Vladimir Hedrih, Pyspost). For students, this is methodically dangerous. Cognitive offloading isn’t that detrimental to our brain functions if it occurs on occasion, but the threshold for concern is when it starts to become an unbreakable habit. What that means for humanity, is essentially that we are slowly becoming used to an easier way of researching, classifying it as real learning, and progressively forgetting what it means to genuinely understand something. Learning is one of the best examples of a positive form of friction in society. It is challenging at times, but beneficial in the long run as it pushes limits in ways we sometimes didn’t even know existed.

AI can be assistive in many ways, and it undoubtedly helps those who don’t always have access to education where there are teachers who prioritize learning that is beneficial to brain function. Undoubtedly, that is why it is so commonly used. Even in the public education system, teachers are beginning to use Artificial Intelligence in their practice. I notice it in my own education; some of my teachers have begun to use Chat-GPT to grade essays or create tests. There have been multiple occasions where my teachers have actually encouraged the usage of AI as a study tool. This isn’t necessarily an issue and in fact, I do use LLMs such as Claude and Chat-GPT to study, especially for niche goals of mine that lack the widespread support system as my general high school curriculum. With that in mind, it is possible that the increased usage of AI in education systems isn’t automatically the fault of the users, but actually due to the system a majority of us have been exposed to. Many researchers commentating on the use of AI have pointed out the fact that the space of academia involves competition that is “insanely fierce,” and therefore visibility and success has mainly become dependent on “cleanly packaged messaging and sheer volume” which can often be more easily achieved with the help of AI (The Atlantic). In addition to this phenomenon, Fairbanks also notices that many writers and communicators are “under increasing pressure to use it, so long as they feel they’re doing so within their profession’s boundaries.” With her credentials as a writer, the growing trends she discusses in her article are hard to ignore. The real question then becomes about what these boundaries consist of, rather than focusing on the ethics of using it to begin with. However, as a society and a culture, we are the ones to determine those boundaries. This may be an issue, because if it is up to us to decide the boundaries and more and more people are using AI, there is really no way to execute a real change. For AI to stop being used in research and academia, it would require everyone to collectively admit that it is unethical and detrimental to human intelligence. In the system of competition, however, it is impossible to be the one to admit this without falling behind. Falling behind would also mean that your words aren’t widely accepted by an audience anymore, and then there would be no point trying to start a movement with no credibility behind it. This tricky paradox is not the fault of human life, but rather of the system itself.

While AI can sometimes be a helpful tool for researchers, it is oftentimes incorrect and includes biases that have a tendency to be harmful to society. It may be regarded as a necessary evil of our society, but the fact is that Chat-GPT, Claude, and Gemini have a tendency to be wrong. Like I have previously asserted, these LLMs include heavy biases based on their trained data, which makes relying on them as a sole source of information a very risky thing to do. A common pattern among these systems is the reinforcement of WHELM ideals, which represents the “western, high-income, educated, liberal and male” perspective (USC Dornsife News). The danger this poses for the well being of already marginalized groups is unmeasurable, and it is important that we remember this when using AI that due to these biases, we cannot take their words as absolute fact. Our individualized feeds on social media and even news sources also make information seem one hundred percent accurate when that is not always the case. This can lead to propaganda at a large scale, and it is scary to consider what this means for the future of certain cultures who aren’t necessarily out of the loop, but simply don’t have access to the truth. Again, one can recognize the lack of friction in the world of academia when the information comes to you rather than through meaningful research. The reality is, though, that this information was designed to grasp your attention and hold it rather than genuinely inform you. In this way, it is not only a lack of friction, but also a lack of authenticity that is becoming more common in spaces of research.

There is also an argument to consider here when regarding the previously established system of education as well. It is reasonable to partially blame the users for the increasing usage of AI and claim that it is a fault of laziness in society, but that is not the sole cause. The more important thing though is that if the system was already corrupt, what will happen when it becomes even more corrupt due to AI usage? Without a proper backbone for society, who will be the ones to decide the ethical standings of our world? Without a culture that prioritizes friction in human life and recognizes that it is beneficial to the well being of society, we have no future to work towards. The future policy makers of our world could be using AI right now, ultimately leading to a lack of critical thinking in the philosophy of a possible society. Will AI be the one to set up the structures of our future? It’s a scary outcome to consider, but not an impossible one. It is perhaps the most important question of modern life and although there are currently no solid solutions, it’s important to begin to consider ways to reverse the current standards of society, starting with a general increase of friction in our everyday lives.

AI’s Contribution to Abundance in Society

In the world we live in now, challenges are not something necessary to human existence. Prior to the creation of technology with the ability to do simple tasks for us, humans had to rely on their own skills in order to create a better life for themselves. Obviously there is a massive list of benefits that come along with the usage of technology—including AI—because otherwise, we wouldn’t be using them nearly as much as we do. The truth is that these technologies do make life easier, but it’s not always as beneficial or sustainable as it may sound. Although life in general may be more convenient and comfortable now, it is fundamentally changing the way we choose to experience our environment, which is particularly challenging to do in the modern world. We can, of course, choose not to focus on AI, but it can be difficult when nearly every app or website features AI “assistive” technology. Rather than simply surviving, we now have the ability to choose the environmental conditions that will help us achieve the most suitable conditions. For many people, this is leading to a lack of general friction in their lives that was unavoidable previously to the creation of AI. I think it is becoming a common theme for this new generation—the ones currently integrating AI into their lives—to hold the expectation that everything in life should come with convenience. This expectation is built on the fact that information is becoming more and more accessible, making the tasks we deem as unimportant just another assignment to cross off a check list. For this reason, I believe that AI contributes greatly to a world full of abundance, but more importantly, it contributes to overindulgence.

From a psychological perspective, AI has led to a disruption in the way our dopamine is regulated, possibly fostering addiction. The ethical concerns of the rising changes in social media is highly concerning, and it is important to understand that this argument isn’t purely based on opinions, but actually backed by scientific research. Pubmed defines the mesolimbic system as the area in our brain responsible for the association between certain stimuli and positive rewards which is what influences our behavior most directly. The main component of this complex system is dopamine, which experts consider to be the core of this process. As it is an influential part of our lives, it is important to understand the possible dangers that are inherently built into this dopamine system. When we find something enjoyable, we have the tendency to recreate this experience in order to feel the same pleasure again and again. The more interesting aspect of this, in my opinion at least, is AI’s contribution to this system. As I have previously mentioned, AI contributes to a more individualized feed and creates a personal experience that has the ability to become an inescapable addiction. Tim Estes defines social media as the “digital heroin” that traps many people in this cycle, and then argues that “new and enhanced AI will become their fentanyl” (Newsweek). Previously, the overuse of social media was the only thing to worry about, but with the creation of AI, there is an entirely new layer of concerns that comes with the overindulgence in technology. The effects of social media—on teenagers specifically—have been studied relentlessly, and it has been proven time and time again that it may seem enjoyable to spend a large percentage of your time on social media, but the effects are highly detrimental. With the addition of AI to this framework, these effects are only going to become more extreme as time goes on. The major reason for this is truly because of the lack of friction that social media and AI breeds. Life becomes easier, dopamine is easier to achieve through low effort tasks, and our typical way of life starts to seem meaningless when we are constantly being stimulated by such extravagant information.

Although it may be an external force causing the decline of friction in human life, that does not mean it is going unnoticed by the vast majority. I think that since the release of LLMs to the general public, people have realized just how detrimental it is to fabricate an understanding with AI rather than doing the work themselves. From the media I personally consume, I have seen a large increase in content that advocates for a rebirth of critical thinking and individual research. When I open my youtube feed, specifically, I am recommended dozens of videos regarding the issues with doomscrolling and how they “got addicted to being disciplined.” What I find ironic about this, though, is that these videos are inherently contributing to the cycle of instant gratification and disrupting dopamine levels. These videos are designed only to engage the viewers and the concept of motivational videos may actually be more dangerous than we realize. The addictive nature of dopamine plays a large role here as well. Researchers Berridge and Kringelbach have defined dopamine as the “wanting” chemical rather than the “having” one. This argument is rooted in the system they define as incentive salience model, which is essentially the craving for dopamine that all humans inherently experience. Their research categorizes dopamine as “evolution’s boldest trick” that is essentially no longer as useful as it once was since we live in “modern environments of abundance” (Berridge and Kringelbach). They specifically define the difference between the ‘wanting’ and the ‘having’ chemicals by stating that “though wanting processes tend to dominate the initial appetitive phase, while liking processes dominate the subsequent consummatory phase that may lead to satiety” (Neuron, vol 86). This claim disputes the popular belief that achieving our goals is what has a tendency to be addictive; the motivation and the pursuit of our goals is what actually releases dopamine, rather than the act of actually succeeding in these goals. For this reason, I believe that the motivational content is essentially just more propaganda that feeds into this constant cycle of addiction and prevents real action from taking place. This further decreases our pursuit of friction in the modern world, but we are able to cover up this unwarranted truth by continuing to plan for changes in our lives only—for many of us—to see little to no action actually taken.

Breaking this cycle may be hard, but it is necessary for human evolution and growth. In the short term, our abundant world and overuse of AI may seem like the easy way out, so why not just continue letting ourselves drown in the cycle? I believe that the long term impacts are much more serious than a mere addiction with the potential to prevent only some people from reaching their goals. It will fundamentally shape our future if we don’t do something about the corrupt reward system otherwise known as our brain chemistry. First off, it will make humans stronger. Hormesis is defined as the process of “small, repeated doses of stress” and actually asserts that it can “make a biological system stronger over time” (PubMed). Seeing as humans are biological systems, this argument factually backs up the claim that friction and hardship are overall beneficial for the human condition, and therefore the longevity of society. Even beyond this simple fact, life becomes more enjoyable when you challenge yourself. Doing the same things day after day has the potential to trap us in a never ending dopamine cycle that can hinder us from reaching our full potential. Elizabeth L. Bjork and Robert Bjork also define an important concept known as “desirable difficulties” which they go on to describe in depth in their research, which even includes specific examples of such difficulties. Their assertion is that rather than trying to simply retain information in your memory, one should “assume that learning requires an active process of interpretation” (Bjork & Bjork). With this argument in mind, along with the fact that learning is the first step to any progress in a civilization, it can be contended that in order to actually grow, we must continue to challenge ourselves. It is more than something that happens at just an individual level, and it will be the responsibility of society as a whole to adopt and then maintain this philosophy on life. Despite that, though, one’s ability to focus on their own personal friction can be mapped as a good first step. The abundance of our world is steering human development towards a path in which learning is no longer even required. In order to prevent this troubling and dangerous trajectory from becoming a reality, we must make the choice for ourselves and realize that learning is truly the only way for long term growth. Friction is not something we should actively avoid when it is so widely beneficial to human development. Instead, we must make the conscious effort to recognize the importance of friction and implement it into our daily lives until eventually it becomes a societal norm.

Conclusion

To some extent, AI does make life easier. It makes daily tasks more convenient, less time consuming, and gives the appearance of performing human responsibilities more efficient than the humans themselves. The usage of AI should not be eradicated, but the copiousness must significantly decrease. Additionally, there are certain aspects of life and even specific fields that would be enhanced if they were more isolated from Artificial Intelligence altogether. LLMs are complicated—and sometimes problematic—systems, but they hold the power to fundamentally reshape the way our society functions. For the majority, what happens regarding large scale AI usage is out of our control. The people who have a say in this will continue to decide for us, but that doesn’t mean we have to be witnesses to our own downfalls. It will most likely be years until we have a clearer idea of what AI will do for humans in the future, and opinions on the subject matter will no doubt evolve over time. At present, all we can do is shift our focus to aspects of our own lives that we can control.

Generally speaking, life is a lot more interesting when we are constantly being faced with experiences we wouldn’t typically seek out. I think it is a general consensus among civilization that diversity is what makes human life so interesting. When we are never faced with challenges or friction, the convenient aspects of our life will begin to become something mundane, and they will no longer be considered an enjoyable privilege. Human life is all about balance, and I think it always has been, simply because it has to be. Our own species still doesn’t completely understand why we think the way we do; it is a mystery why humans are conscious on a level that other beings are not. On that note, the way we reach homeostatic balance is different from any other animal, and I think that for humans in particular, friction is an essential part in the path to reach homeostasis. Ultimately, I believe the saving factor of this entire chain reaction is to keep in mind the three crucial things humans have that AI does not: consciousness, creativity, and emotional capacity. Though AI can mimic these things, it will never have the same meaning behind it that it does when produced by a human. Cohesively, we must look inward and remember the lost value of human friction and hardship. Looking inward is about realizing what real change matters at a large scale and having the ability to consider other beings rather than just yourself. It is in no way helpful to be “morally elite” by refusing to use AI. Instead, it’s important to look at the possible long-term effects and focus on a rebirth of critical thinking. That way, we can critically think about what the overuse of AI means for the future of society. It really does have the potential to create a morally and systematically dysfunctional one—more than the one we already reside in—and only we have the ability to prevent that from happening.

Works Cited

Berridge, Kent C., and Morten L. Kringelbach. “Pleasure Systems in the brain.” Neuron, vol. 86, no. 3, May 2015, pp. 646–664, https://doi.org/10.1016/j.neuron.2015.02.018.

Bjork, Elizabeth Ligon, and Robert A. Bjork. “Chapter Learning Elizabeth L. Bjork and Robert Bjork.” Psychology and the Real World, UCLA, 20 Nov. 2009, bjorklab.psych.ucla.edu/wp-content/uploads/sites/13/2016/04/EBjork_RBjork_2011.pdf.

De, Debasmita, et al. “Social media algorithms and teen addiction: Neurophysiological impact and ethical considerations.” Cureus, 8 Jan. 2025, https://doi.org/10.7759/cureus.77145.

Estes, Tim. “Ai Will Be Youth’s Fentanyl.” Newsweek, Newsweek, 30 Apr. 2024, www.newsweek.com/if-social-media-digital-heroin-todays-youth-ai-will-their-fentanyl-opinion-1895335.

Fairbanks, Eve. “The Biggest Tell That Something Was Written by Ai.” The Atlantic, Atlantic Media Company, 29 May 2026, www.theatlantic.com/technology/2026/05/how-to-tell-ai-writing/687345/.

Hedrih, Vladimir. “Study Finds CHATGPT Eases Students’ Cognitive Load, but at the Expense of Critical Thinking.” PsyPost, PsyPost Media Inc., 17 Sept. 2024, www.psypost.org/study-finds-chatgpt-eases-students-cognitive-load-but-at-the-expense-of-critical-thinking/?utm_source=chatgpt.com.

Joy, Darrin, and Zhivar Sourati. “Ai Is Changing More than Writing - It May Be Shaping Our Worldview.” USC Dornsife News, USC Viterbi School of Engineering, 2 Apr. 2026, dornsife.usc.edu/news/stories/ai-may-promote-cultural-homogenization/.

Klebanov, Sam, and Morning Brew. “Linguists Say CHATGPT Is Now Influencing How Humans Write and Speak.” Fortune, Fortune, 30 June 2025, fortune.com/2025/06/30/linguists-chatgpt-influencing-how-humans-write-speak/.

Kriss, Sam. “Why Does A.I. Write like … That? - The New York Times.” The New York Times Magazine, The New York Times Company, 3 Dec. 2025, www.nytimes.com/2025/12/03/magazine/chatbot-writing-style.html.

Mattson, Mark P. “Hormesis defined.” Ageing Research Reviews, vol. 7, no. 1, 7 Jan. 2008, pp. 1–7, https://doi.org/10.1016/j.arr.2007.08.007.

Natalie Wiesner, "Minds, Machines and Medicine: How Artificial Intelligence is Reshaping Surgery, Diagnostic, and the Future of Healthcare"

Artificial intelligence has moved from Science Fiction into the operating room, the clinic, and the pharmacy lab. The impact of AI on medicine has been felt in less than two decades. In this regard, there have been the creation of robotic surgical systems that are assisted by machine learning algorithms, the development of diagnostic tools that can spot cancers that cannot be detected with the naked eye, and ambient recording technology that documents conversations between doctors and patients without their knowledge. These changes in medicine represent some of the promises of artificial intelligence in healthcare including quicker diagnosis, reduced human error, more tailored treatments, and new pharmaceutical drugs that would not have been discovered otherwise. At the same time these developments present some important concerns because AI moves from being the mere helper of the clinicians to the decider that determines the life or death of a patient. 

The surgical suite is one of the toughest environments for doctors in the field of medicine because they need to take into account a huge amount of information during operation and work precisely with the ability to react instantly to unexpected challenges. One of the most well-known AI driven surgical robots used by medical professionals today is the da Vinci surgical system. Such technology allows performing surgical procedures in a way which is not possible when a doctor uses only his hands. It allows translating the movements of a surgeon into more accurate movements, eliminating any tremors, increasing magnification and the ability to operate through incisions of just several millimeters. This is just some of what these devices have accomplished, and the result has been a significant decrease in blood loss, shorter hospital stays, decreased post-procedure pain, and faster recovery for patients compared to traditional open surgeries. Yet, the application of AI in surgical and clinical decision making has expanded its scope from assisting a surgeon physically to making or at least significantly impacting life or death decisions and this trend poses serious ethical questions. Now predictive AI models are employed to estimate risk of patient deterioration, calculate the probability of developing sepsis and identify candidates for early intervention in intensive care units. In some hospitals, this is happening on an ongoing basis as AI continuously processes the stream of psychological data coming from bedside monitors and electronic medical records of the patient and procedures risk estimates which the nursing staff and physicians are supposed to follow up. While such solutions are referred to as decision support tools, in reality, there is not always enough time in busy clinical settings for critical evaluation of the algorithm's output. In case when a patient is marked by an AI model as high risk, the actual outcome is the same as the algorithm recommendation becoming the decision. It becomes all the more crucial when AI becomes part of the emergency triage systems, models for allocating organ transplants, and guidelines for end of life care. In each of these scenarios, an algorithm based on historical data and probability is playing a role in deciding or, according to come deciding entirely who will get a scarce good, who will be upgraded to intensive care, and who won’t. The claim is that such algorithms help alleviate the inevitable biases and inconsistencies of human judgement, especially during times of stress and fatigue. This is a valid point: it has been demonstrated many times that doctors' decisions show considerable variation according to completely arbitrary criteria. Algorithmic systems are not bias-free, however; they just process bias differently. In this case of training AI models based on the historic medical data, they might reproduce and even exaggerate the existing discrimination due to the fact that the historic reality was shaped by the unequal healthcare system. It has been found through landmark research that they risk stratification algorithms commonly used in American hospitals systematically under-assessed the illness in black patients leading to fewer referrals for them into intensive care programs, not because they were healthier than others, but due to the fact that their healthcare needs were not assessed correctly as the algorithm relied in the healthcare spending, while Black people historically lacked the same access to healthcare. This example demonstrates the crucial aspect; an AI system cannot be considered safe enough to make life-and-death decisions without careful analysis of its assumptions, data, and outputs.

Perhaps one area where the impact of artificial intelligence in the domain of medicine is felt most tangibly is medical imaging and diagnostics. This is because radiology, pathology, dermatology, and ophthalmology are specialties whose core responsibility involves reviewing visual information - X-ray’s, CT scan results, MRI’s Tissue samples, skin images, and retinal images - and determining the presence of abnormalities which signify illness. This is the type of pattern recognition that deep learning excels at. With the ability to analyze hundreds of thousands, if not millions, of labeled images, AI programs will be able to determine specific characteristics of disease with accuracy far superior to that of human analysis. The diagnostic power of AI in radiology is one of the most thoroughly investigated areas and has proved itself through several studies and applications. The models created by Google Health, Zebra Medical Vision, have proven their ability to find pneumothorax, pulmonary nodules, vertebral fractures, and intracranial bleeding as effectively as or even better than trained radiologists in controlled conditions. In breast cancer screening specifically. There is considerable progress as well. As reported in the article in Nature in 2020, the machine was able to detect breast cancer more accurately from mammograms than six radiologists, minimizing false positives and negatives and being better at detecting breast cancer on its own or when combined with a human reader. Given that early diagnosis is the only critical factor in the disease, the significance of this result can be enormous. In pathology, the use of AI for examining digitized tissue samples at extremely fine detail is being employed. AI can help study the cellular structure of biopsy samples by counting mitoses, recognizing tumor-infiltrating lymphocytes, and characterizing cancer types. All of this would take a much longer time for an expert in pathology to do. SOme algorithms have shown their capability to predict molecular characteristics of tumors - including mutations in specific genes useful for choosing the appropriate treatment options - from routine histopathological images, without any need for costly molecular analysis required for this purpose. In poor and developing nations where pathologists are rare and molecular analysis is not available, AI-powered pathology tools can provide patients with the information on their cancer which would not be otherwise accessible. Still, AI functions better when used as an additional pair of eyes as opposed to being the only set. This is because the system is unpredictable in its ability to recognize images that are different from those used in the training. Human experts have experience in clinical judgement that is yet to be stimulated by AI. 

Doctors spend almost two hours writing documents for each hour of direct contact with patients, and it became a main driver for the burnout epidemic in medicine. Ambient scribing solutions based on artificial intelligence like Nuance, Abridge, and Suki Al can deal with this problem. They listen to the conversation between doctors and patients and prepare clinical documentation, diagnoses, and electronic health records automatically. After that, doctors check it and approve it, but they do not need to waste their time transcribing. Healthcare providers save a lot of time on documenting outside working hours, while patients feel more attended when their doctors look at them and not their computers during a consultation. However, some issues related to privacy and accuracy of generated data should be considered. Patients should give their explicit consent to ambient recording, and healthcare providers should proofread the information provided by AI solutions.Nonetheless, ambient documentation could become one of the most impactful uses of artificial intelligence in medicine not due to technology, but due to the fact that it saves doctors attention and time. 

The development process of a drug, from its discovery till approval by regulatory agencies, can take more than ten years and cost over two billion dollars, with a success rate of less than ten percent. AI is making changes to this process. Deep learning algorithms are able to assess the behavior of potential drug molecules inside the human body, assess their toxicity, and find those compounds which would be very difficult for any group of human researchers to find. DeepMind’s AlphaFold, which had predicted three-dimensional structure of almost all proteins, has paved the way for an entirely new world for structure-based drug design. Several AI-generated drugs have already reached the clinical trials stage. The companies such as Insillico Medicine and Exscientia have designed molecules in much shorter time compared to traditional methods. This indicated that AI may help discover treatments of the diseases that have remained out of reach till now. 

The impact of AI on medicine does not happen at once but as an ongoing trend in the field of surgery, diagnostics, documentation, and drug discovery. In all of these areas, the potential for improvement is high, but the limitations are also considerable. The physician who evaluates the suggestion provided by AI technology, endorses notes made by AI algorithms, or appraises the prospects of a new drug developed by artificial intelligence remains the crucial safeguard in the process. The future of medicine is not the scenario of substitution of a doctor by some machine but rather the scenario of collaboration of human and artificial intelligence. This partnership must be built on solid foundations of ethics and high responsibility for patient safety.

Works Cited

DiMasi, Joseph A., et al. “Innovtion in the pharmaceutical industry: New Estimates of R&D Costs.”Journal of Health Economics, vol 47, 2016, pp. 20-33.

Hashimoto, Daniel A., et al. “Artificial Intelligence in surgery: Promises and perils.” Annals of Surgery, vol. 268, no 1, 2018, pp. 70-76.

Jumper, John, et al. “Highly Accurate Protein structure prediction with AlphaFold.” Nature, vol. 596, 2021, pp. 583-589

Kather, Jakob Nikolas, et al. “Pan-cancer Image Based deception of clinically actionable Genetic Alterations." Nature Cancer, vol. 1, 2020, pp. 789-799.

McKinney, Scott Mayer, et al. “International Evaluation of an AI System for Breast Cancer Screening.” Nature, vol. 577, 2020, pp.89-94.

Nath, Bhavanna et al. “Ambient Clinical Intelligence: The promise of Reducing Clinician Burnout Through AI-Enabled Documentation.” npj Digital Medicine, vol. 5, 2022, article 167.

Obermeyer, Ziad, and Ezekiel J Emanuel. “Predictong the future - Big Data, Machine Learning, and Clinical Medicine.”New England Journal Of Medicine, vol. 375, no.13,2016,pp.1216-1219. 

Noah Ivanchikova-Crouse, "Balancing Innovation with Regulation"

The rise of AI has resulted in its incorporation all throughout society, being used in workplaces, schools, and governments. As it continues to advance, its place industries and governments around the world continue to grow. This use of a new technology has caught the world off guard, creating a plethora of issues that need addressing, such as privacy and copyright concerns. These issues remain very difficult for governments to successfully regulate, as the growth of AI is often faster than the institutions developing certain frameworks. However, despite these risks associated with AI, it can also grant serious economic benefits to states who successfully integrate it into their systems, creating an interesting dilemma for governments needed to weigh the economic benefits against the potential risks. As a result, nations must choose to balance innovation with regulation.

The new appearance of AI within our society has led to a number of difficulties that need to be addressed. One of the most highly discussed topics among these difficulties is AI’s place is the current landscape for copyrighted materials. The modern AI systems of today are trained using large amounts of data sets, that all help the system develop. The main legal question this presents is if AI companies should be allowed to train their chatbots on copyrighted materials. Many AI producers argue that they should be able to train on copyrighted works as  AI isn’t fully reproducing the original works, but using the original to learn patterns. On the other hand, many producers and artists say their content is being used without their permission, and as a result they should be given compensation.  This conflict has led to many creators becoming upset and suing AI companies for using their material. Many of the major legal cases are leaning towards the side of the creator, against the AI companies. In 2020, through a court case located in Los Angeles, Thomson Reuters sued Ross Intelligence, an AI company, for training their model on Reuters’ content without permission. The judge decided that Ross Intelligence was not permitted to use Reuters content, under copyright law. This was one of the earliest lawsuits surrounding AI copyright, and set the stage for how future cases were going to be settled. Now, many more creators are favored over AI companies, potentially forcing future AI companies to obtain licenses for their material, a telling sign in how AI training will have to change.

Another one of the more concerning challenges associated with artificial intelligence is the possibility of misinformation, and the scary capabilities of deepfakes. AI can produce information quickly, cheaply, and at a very large scale. This can lead to the production of a large amount of information, and if this production continues, it could overtake the amount of trustworthy information very easily. If this happens, it would essentially undermine any trust the public has in news sources, with any source being potentially AI. When AI becomes increasingly developed, it would be impossible to tell the difference between regular information and AI, creating a very harsh future for how people receive their information. The possibility of deepfakes is another scary future, being able to reproduce basically any form of content. They could be used for misinformation, financial fraud, election interference, or many other malicious ways. This would severely undermine a lot of the trust in public information organizations, as well produce severe risk through further development of AI.

In addition to concerns about copyright and misinformation, AI also creates serious threats to our data and privacy. AI affects our privacy in a surprising amount of ways, from the use in training, fueling consumerism, and even the development of facial recognition. As a result of AI requiring so much data in order to train effectively, much of this data comes from social media platforms or public records. While this data is technically public, it still arouses many concerns of users worried about their privacy among these enormous systems. Despite these concerns, much of the training on social media is fully permitted, as a majority of social media companies automatically require users to allow their information to be collected, as it is a default listed in their terms of service.

Another threat that AI presents to privacy is the new development of facial recognition technology. These technologies are able to identify individuals based on photographs, and videos. This allows for people to be tracked in public spaces, and assistance in law enforcement. This results in an interesting dilemma between the value of privacy and the possibility of improving public safety. Electric Frontier Foundation (EFF) has created a campaign to try and remove facial recognition from the government entirely. They argue that it threatens constitutional rights, by taking away the right to exist in public anonymously. EFF also has done research showcasing the inaccuracies of facial recognition. Their studies have revealed that there are high error rates when identifying women, young people, and people of color. They also mention the risks of facial security, as unlike your credit card, you cannot change your face if the information is breached. As a consequence of this outcry, many cities, such as San Francisco and Boston, have issued bans preventing the use of facial recognition within their law enforcement. The movement against this technology is growing rapidly, and groups will continue to fight until it is completely banned.

Despite these risks, governments are still choosing to incorporate AI into their systems for a number of different reasons. AI has a lot of potential to serve within the government, as advanced versions of this technology can do a number of things. One of the most immediate benefits that AI could have is the ability to improve government efficiency. Governments being able to automate simple tasks through AI could lead to much faster creation, implementation, and enforcement of certain government policies. Advanced AI systems could also prove to be a vital contributor to national security. AI systems would be much more efficient at being able to identify potential threats or suspicious activities that could cause harm. National security is a major priority for many governments, so being able to institute AI in this way could serve as a very strong compeller for development. Another way AI could serve the government is through military means. AI has the potential to drastically increase military capabilities by improving decision making, efficiency, and weapons. Modern militaries generate very large amounts of data through satellites, drones, and radar systems. AI being able to analyze this data more efficiently could allow military teams to be better at identifying threats, tracking enemies, and making decisions. AI also has a place in advancing current military technologies. AI operated technologies like drones and missiles have shown that they could be very effective in military combat when developed further. AI’s place in military technology offers another incentive for governments to continue to develop this technology.

The development of AI can help a country through other ways, and also bring new powerful companies and institutions, providing significant benefits to the countries that have more advanced artificial intelligence. Successful AI companies contribute to a nation’s economy in a number of ways. The research, data, and resources needed to develop AI results in a significant boost in GDP. As the companies expand, they also produce billions of dollars worth of investments. This is comparable to past innovations, such as the internet. The countries that become leaders in this development will experience a significant increase in economic growth. The development of AI could also contribute to innovation and scientific advancement. AI could contribute to medicine, robotics, biotechnology, and many other fields of innovation. This potential to expand certain aspects of scientific development, as well as potential economic growth, serve as a strong competitive advantage towards governments that choose to continue developing AI systems, rather than prohibiting it. 

Nations around the world are taking very different actions when determining AI system regulation. Their decisions reflect how they value economic growth, compared to Currently, the European Union act has taken the most comprehensive approach to AI regulation. This document's purpose is to ensure the safety of AI utilization, specifically aiming to make AI usage transparent, traceable, and respectful. Everything established in the document is supported through strict enforcement by the law. This document organizes AI systems into 4 different tiers, based on the potential harm they could impose. The first tier is systems that have too much risk to allow. These systems are strictly prohibited as they are deemed a threat to society.  Examples include cognitive manipulation or social scoring. Social scoring is using AI to analyze public behavior, and assign values to individuals based on their behavior, in order to judge their trustworthiness or value. The second tier is systems with high risk, which are systems that could potentially have significant impact on human rights and ordinary lives. Examples of this include implementation in education, healthcare, or employment. These are strictly monitored, and must go through rigorous testing before being implemented. They must maintain data logging, have high security, and be registered in the EU database. The third tier is limited risk. These are systems that could very well influence daily life or media, but wouldn’t have much possibility to affect fundamental institutions. Examples of these include AI chatbots, or manipulation devices, such as deep fakes. Users must have complete transparency when using these systems, in order for users to remain fully informed that they are interacting with AI produced media. The last tier is minimal risks, which include video games or spam filters. These face no structural regulation.

This is very different than how the US is approaching regulation, opting for much more decentralized regulation. The US doesn’t utilize broad, overall federal legislation. Instead, each state develops their own distinct version of AI regulations. This creates a framework of regulation that is unclear, but also succeeds in facilitating AI growth over other countries. The U.S. government is cautious about implementing any AI policies, as they see it as a source for economic growth. Therefore, many of the restrictions the US implements focus on regulating harmful outcomes, rather than AI development itself. The approach is what allows the US to remain leaders in the global AI industry, however, critics argue that they should be prioritizing the creation of a comprehensive framework that encompasses the whole nation, sufficiently addressing important issues. As AI continues to expand, more pressure is being placed on congress to establish clear national standards. 

China, similar to the EU, has developed a more centralized approach to AI regulation, but similar to the US, they also see AI development as vital for their economic growth. Under China’s regulations, companies must conduct security reviews, monitor content, and make sure their product complies with government standards. These regulations are enforced by law enforcing agencies, such as the Cyberspace Administration of China. The government of China does still hold the development of AI as very important to their national identity, as they have incorporated AI into many of their development plans. According to the Library of Congress, the government strictly prohibits any content that threatens state power, national security, or any undermines social stability. Developers also must undergo rigorous security assessments in order to qualify for algorithms. They also ensure that all the data used to train AI is legally acquired needing to respect property rights. These regulations are only placed on companies generating AI services for the public, and not companies developing AI for internal use. This successfully balances the control needed without the cost of the economic support AI systems provide. 

As artificial intelligence has rapidly transformed society, it has created significant legal challenges for governments across the world. Offering a new set of challenges, law makers are having trouble addressing the concerns surrounding copyright, privacy, and misinformation. The governments around the world are choosing to navigate these risks in order to profit from the benefits that AI presents them. The governments that can successfully encourage future AI development will be able to fully profit off of its transformative potential, while the ones that choose to address the concerns will not be harmed by any of the risks. This is leading to a large discrepancy into how different sectors are approaching AI development, with some choosing to value the potential growth over the risks, while others are cautious of the harm that AI could cause.

Works Cited

“About Face.” Electronic Frontier Foundation, 8 Nov. 2019, www.eff.org/aboutface. Accessed 19 June 2026.

“Artificial Intelligence Act.” Consilium, 2020, www.consilium.europa.eu/en/policies/artificial-intelligence-act/. Accessed 19 June 2026.

“Artificial Intelligence Compliance Plan.” Federal Trade Commission, 2 Oct. 2024, www.ftc.gov/ai. Accessed 29 June 2026.

“China: Generative AI Measures Finalized.” The Library of Congress, 2023, www.loc.gov/item/global-legal-monitor/2023-07-18/china-generative-ai-measures-finalized. Accessed 19 June 2026.

Choi, Abigail. “Public Records Act Meets AI: What Recent Records Requests Reveal about ChatGPT Use in Government.” Liebert Cassidy Whitmore, 12 May 2026, www.lcwlegal.com/news/public-records-act-meets-ai-what-recent-records-requests-reveal-about-chatgpt-use-in-government/. Accessed 19 June 2026.

Harris, Laurie. “Regulating Artificial Intelligence: U.S. And International Approaches and Considerations for Congress.” Congress.gov, 2025, www.congress.gov/crs-product/R48555. Accessed 19 June 2026.

Hengesbach, Ellen . “How to Stop Your Data from Being Used for AI Training.” U.S. PIRG Education Fund, 12 Nov. 2024, pirg.org/edfund/resources/how-to-stop-your-data-from-being-used-for-ai-training/. Accessed 19 June 2026.

“Organization.” Ai.mil, Jan. 2021, www.ai.mil/About/Organization/. Accessed 19 June 2026.

PARVINI, SARAH. “Thomson Reuters Scores Early Win in AI Copyright Battles in the US.” AP News, 12 Feb. 2025, apnews.com/article/ai-artificial-intelligence-reuters-4a127c5b7e8bb76c84499fe12ad643c8. Accessed 19 June 2026.

“The 2026 AI Index Report.” Stanford.edu, 2026, hai.stanford.edu/ai-index/2026-ai-index-report. Accessed 19 June 2026.

Angelina Castillo, "The Ethics of Artificial Intelligence: How AI is Reshaping Human Life"

Introduction

Artificial intelligence has rapidly become one of the most influential technologies of the twenty-first century. From recommendation algorithms on social media to generative tools capable of producing text, images, and videos, AI is increasingly integrated into everyday life. Supporters argue that AI can improve efficiency, expand access to information, and solve complex problems. Critics however warn that AI presents significant ethical challenges involving privacy, bias, labor, and human autonomy. As AI becomes more powerful, society faces an important question: how can humans ensure that artificial intelligence serves human interests rather than undermining them? While AI offers significant benefits, its growing influence on decision-making, employment, privacy, and social interactions demonstrates the need for ethical oversight and human centered development. One of the most important ethical concerns surrounding AI is the way it influences human decision making. Artificial intelligence systems are often presented as objective and rational because they rely on data and algorithms. However, AI systems are only as reliable as the data they are trained on. If historical data contains biases, those biases can become embedded in the technology itself. This creates the illusion that AI-generated decisions are neutral when, in reality, they may reinforce existing inequalities. Philosophy Tube’s video “AI Is an Ethical Nightmare” highlights this issue by examining real-world examples of algorithms used in hiring practices, criminal justice systems, and financial institutions. The video argues that many AI systems inherit biases from the societies that create them leading to discriminatory outcomes that can affect people’s lives in profound ways. For example, hiring algorithms trained on historical employment data may favor applicants who resemble previous successful candidates potentially disadvantageous to women or minority groups. Rather than eliminating human bias AI can automate and amplify it. The ethical implications of biased AI extend beyond individual cases. When algorithms make decisions about employment, healthcare, housing, or criminal sentencing, they influence opportunities and life outcomes for millions of people. Because these systems often operate behind complex technical processes, affected individuals may not understand why a decision was made or how to challenge it. This lack of transparency raises concerns about fairness and accountability. If society increasingly relies on AI to make important decisions, ensuring transparency and oversight becomes essential.

 Another major ethical issue is the impact of AI on employment and economic inequality. Throughout history, technological innovations have transformed labor markets, but AI has the potential to automate both physical and cognitive tasks. Jobs that once required human judgment can now be performed by algorithms capable of processing large amounts of information at unprecedented speeds. Supporters of AI often argue that automation will create new opportunities and increase productivity. Dominic Burbidge, in “AI Ethics Is Simpler Than You Think,” suggests that AI should be understood primarily as a tool rather than as an all-powerful force that will replace humanity. He argues that AI’s strengths lie in specialized tasks such as pattern recognition and optimized responses rather than general human intelligence. According to Burbidge, ethical AI development requires focusing on practical applications and ensuring that technology serves specific human needs rather than imagining AI as a replacement for people. Although Burbidge’s perspective emphasizes the limitations of AI, concerns about job displacement remain significant. Many industries are already adopting AI-powered systems to reduce labor costs and increase efficiency. Customer service representatives, data analysts, content creators, and even some legal professionals face increasing competition from automated tools. Workers whose jobs are partially automated may experience reduced wages or fewer employment opportunities. The benefits of AI often accrue to large technology companies and investors, while workers bear the costs of economic disruption. Philosophy Tube expands on this concern by emphasizing that AI systems do not operate independently of social and economic structures. The video argues that discussions about AI ethics cannot be separated from questions of power, labor, and inequality. AI development relies on extensive human labor, including data collection, content moderation, and system training. These workers are often underpaid and invisible to the public despite being essential to the functioning of AI technologies. Ethical discussions about AI therefore require attention not only to algorithms but also to the people who build and maintain them. 

Privacy represents another critical ethical challenge. Modern AI systems depend on enormous quantities of data. Every online search, social media interaction, purchase, and location update contributes to a growing digital profile. Companies use this information to train algorithms, personalize content, and predict user behavior. The collection of personal data raises concerns about surveillance and consent. Many users do not fully understand how their data is gathered, stored, or used. AI systems can analyze information at a scale far beyond human capability, identifying patterns that reveal intimate details about individuals’ preferences, habits, and beliefs. While these capabilities may improve services and convenience, they also create opportunities for misuse. For example, targeted advertising systems use AI to predict consumer behavior and influence purchasing decisions. Political campaigns have similarly used data-driven algorithms to tailor messages to specific groups of voters. These practices blur the line between persuasion and manipulation. If AI systems can accurately predict human behavior, they may also be capable of shaping it in ways that undermine individual autonomy. Privacy concerns become even more significant when governments use AI technologies for surveillance purposes. Facial recognition systems, predictive policing tools, and data-monitoring programs have expanded in many parts of the world. Although proponents argue that such technologies improve public safety, critics warn that they threaten civil liberties and disproportionately affect marginalized communities. Ethical AI development must therefore balance technological capabilities with respect for privacy and human rights.

In addition to privacy concerns, AI has significant effects on human relationships and social behavior. Social media platforms increasingly rely on AI algorithms to determine what content users see. These algorithms are designed to maximize engagement by showing content that attracts attention and encourages interaction. While this approach may increase user activity, it can also contribute to misinformation, polarization, and addiction. Algorithms often prioritize emotionally charged content because it generates stronger reactions. As a result, users may be exposed to sensationalized or misleading information that reinforces existing beliefs. This process can create echo chambers in which individuals encounter only perspectives that align with their own views. The spread of generative AI technologies introduces additional challenges. AI systems can now produce realistic images, videos, and text that are difficult to distinguish from human-created content. Deepfakes and AI-generated misinformation have the potential to undermine trust in media, institutions, and public discourse. If people can no longer determine what is real, the foundations of informed democratic decision-making may be weakened. Philosophy Tube argues that ethical discussions about AI should focus not only on hypothetical future dangers but also on the real harms occurring today. Rather than imagining a distant future dominated by superintelligent machines, the video emphasizes the immediate social consequences of existing technologies. Issues such as algorithmic discrimination, labor exploitation, and misinformation already affect millions of people. Ethical responsibility therefore requires addressing present challenges rather than focusing exclusively on speculative scenarios.  Despite these concerns, AI also offers substantial benefits that should not be ignored. In healthcare, AI systems assist doctors in diagnosing diseases, analyzing medical images, and identifying treatment options. In education, AI-powered tools provide personalized learning experiences that adapt to individual students’ needs. Researchers use AI to analyze scientific data, model climate patterns, and accelerate discoveries that could improve quality of life. These applications demonstrate that AI itself is not inherently ethical or unethical. Rather, the ethical implications depend on how the technology is designed, implemented, and governed. Burbidge argues that AI should be viewed as a limited instrument that requires human judgment and domain-specific expertise. He emphasizes the importance of developing AI systems in close collaboration with the people who will use them. According to this perspective, ethical AI emerges not from abstract rules alone but from practical engagement with real human needs and values. This emphasis on human-centered design offers an important framework for addressing AI ethics. Developers, policymakers, and users must work together to establish standards that prioritize transparency, fairness, and accountability. Ethical guidelines should require organizations to explain how AI systems make decisions and provide mechanisms for individuals to challenge harmful outcomes. Governments should implement regulations that protect privacy and prevent discriminatory practices while still encouraging innovation. Education also plays a critical role in ethical AI development. As AI becomes more integrated into society, citizens need a basic understanding of how these systems function and what limitations they possess. Public awareness can help individuals make informed decisions about technology use and hold organizations accountable for unethical practices.

Ultimately, the ethics of artificial intelligence is not simply a technical problem but a human one. AI reflects the values, priorities, and assumptions of the societies that create it. The choices made by developers, corporations, and governments will determine whether AI contributes to human flourishing or exacerbates existing inequalities. Ethical AI requires more than advanced technology; it requires a commitment to justice, transparency, and respect for human dignity. Artificial intelligence is transforming the modern world at an extraordinary pace. Its effects on decision-making, employment, privacy, and social relationships demonstrate both its potential and its risks. As Dominic Burbidge argues, AI should be understood as a tool that serves human purposes rather than as an autonomous force destined to replace humanity. At the same time, Philosophy Tube’s analysis reminds audiences that AI technologies already create ethical challenges involving bias, labor, surveillance, and misinformation. Together, these perspectives highlight the importance of human-centered approaches to AI development. By prioritizing accountability, transparency, and fairness, society can harness the benefits of AI while minimizing its harms. The future of artificial intelligence will ultimately depend not on machines themselves but on the ethical choices humans make in designing and using them.

Works Cited

Scholarly Source:

Burbidge, Dominic. “AI Ethics Is Simpler Than You Think.” The New Atlantis, no. 82, 2025, pp. 80–89. JSTOR, https://www.jstor.org/stable/27403536. Accessed 14 June 2026.

Popular Source:

https://youtu.be/AaU6tI2pb3M?si=72L5RuaSJP-j2ubb

Want to Keep Up to Date with Our Reading List?

See below for each week's readings and join the conversation!

Weeks 37-38 - No assigned readings

Week 36

1. Michael Gerlich, “AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking”

2. Nataliya Kosmyna, “Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using AI” pages 1-4, 21, 

3. Jinrui Tan, “Learners’ AI dependence and critical thinking: The psychological mechanism of fatigue and the social buffering role of AI literacy”

4. James O’Sullivan, “Stylometric comparisons of human versus AI-generated creative writing”

5. Particle6, “Hi, I’m Tilly Norwood!”

6. Masahiro Mori, “The Uncanny Valley: The Original Essay by Masahiro Mori” 

Week 35

1. Jacqueline Fendy, “Vibocracy and the Collapse of Shared Reality”

2. Jodi Kantor, “The Inside Story of Five Days That Remade the Supreme Court”

3. Tony Rehagen, “Welcome to Post-Truth America”

4. Maxi Heitmayer, “The Second Wave of Attention Economics. Attention as a Universal Symbolic Currency on Social Media and beyond” 

Week 34

1. Wiki, Utopia

2. Thomas More, Utopia selections

3. Jamie Dimon, “Letter to Shareholders 2026”

4. US Govt, “National Security Strategy November 2025 Policy Document” (from last week’s readings!)

5. Alex Karp, The Technological Republic in Brief (from X)

6. Talmon Smith, “The Greatest Wealth Transfer in History Is Here, With Familiar (Rich) Winners”

Week 33

1. Alexis Madrigal, “The Energy in Things”

2. Eve Warburton, “Nationalist enclaves: Industrialising the critical mineral boom in Indonesia”

3. Nisha Talagala, “Data as The New Oil Is Not Enough: Four Principles For Avoiding Data Fires”

4. Juana Summers, “Dirty nickel: The health costs of mining in Indonesia”

5. Lawrence Wright, “Lithium Dreams”

6. Greg Rosalsky, “Why the AI world is suddenly obsessed with a 160-year-old economics paradox”

7. Camilla Domonoske, “Their batteries hurt the environment, but EVs still beat gas cars. Here's why.”

8. Rithwik Kalale, “Lithium mining for EVs: how sustainable is it?”

9. Soumya Karlamangla, “The California Lake Billed as the ‘Saudi Arabia of Lithium’”

10. Image, “This is a silicon wafer” 

Week 32

1. 1. Wiki, “Mother Nature”

   2. Ken Silverstein, “America’s AI Boom Is Running Into An Unplanned Water Problem”

   3. Elizabeth Kolbert, “The E.P.A. vs. the Environment”

   4. Adam Zewe, “Explained: Generative AI’s environmental impact”

   5. Shaolei Ren, “The Uneven Distribution of AI’s Environmental Impacts”

   6. Carmen Gonzalez, “Environmental Justice, Human Rights, and the Global South Environmental Justice, Human Rights, and the Global South”

   7. Megan Mastrola, “How AI Can Help Combat Climate Change” 

Week 31

1. Nathan Heller, “Is the Gig Economy Working?”

2. Matt Shumer, “Something Big is Happening”

3. Kashmir Hill, “Chatbots Can Go Into a Delusional Spiral. Here’s How It Happens.”

4. Jennifer Devries, “How Bad Are A.I. Delusions? We Asked People Treating Them.”

5. Rhitu Chatterjee, “Their teenage sons died by suicide. Now, they are sounding an alarm about AI chatbots”

6. Stuart Heritage, “‘I felt pure, unconditional love’: the people who marry their AI chatbots

7. Shannon Bond, “AI chatbots upended their lives. Now they're finding support from each other”

8. Johanna Costigan, “China's AI Boyfriend Business is Taking On a Life of Its Own”

9. Miles Klee, “This Spiral-Obsessed AI ‘Cult’ Spreads Mystical Delusions Through Chatbots” 

Optional:

1. Andrew Clark, “The Ability of AI Therapy Bots to Set Limits With Distressed Adolescents: Simulation-Based Comparison Study”

2. Juliet Schor, “Dependence and precarity in the platform economy” 

Week 30

10. Wiki, “Satoshi Nakamoto”

11. John Carreyrou, “My Quest to Solve Bitcoin’s Great Mystery”

12. Satoshi Nakamoto, “Bitcoin: A Peer-to-Peer Electronic Cash System”

13. TechScope, “Meme Coin Mania and Mayhem: Navigating the Blind Spots”

14. CoinMemeCap, “Top Memes Tokens by Market Capitalization”

15. Florida State, “Ponzi Schemes”

16. Ganesh Sitaraman, “We Must Prepare for an AI Bubble Now”

17. Michał Klincewicz et al., “Slopaganda: The interaction between propaganda and generative AI”

18. Kyle Chayka, “The Team Behind a Pro-Iran, Lego-Themed Viral-Video Campaign”

Week 29

Week 28

19. Independent Reading Selections (second half)

Week 27

1. Independent Reading Selections (first half)

Weeks 25/26

Does AI Have a Soul? Ethical AI, Palantir, Anthropic, and the US Government

1. Berber Jin, “Meet the One Woman Anthropic Trusts to Teach AI Morals”

2. Image, “Should I Walk to the Car Wash?”

3. Matteo Wong, “Drink Whole Milk, Eat Red Meat, and Use ChatGPT”

4. Amrith Ramkumar, “Pentagon Used Anthropic’s Claude in Maduro Venezuela Raid”

5. Claude, “Claude’s Constitution”

6. Socrates, from Phaedrus on the soul

7. Various, Palantir Readings

1. 1. Mona Khalil, “Palantir, Seemingly Everywhere All at Once”

   2. Steven Hubbard, “ICE to Use ImmigrationOS by Palantir, a New AI System, to Track Immigrants’ Movements”

   3. Alex Karp, “Letter to Shareholders February 2026”

   4. Gideon Lewis-Kraus, “The Palantir Guide to Saving America’s Soul”

Week 23 

How Does it Work? The Basics of AI Functionality

1. Janelle Shane, You Look Like a Thing and I Love You: How AI Works and Why It’s Making the World a Weirder Place chapters 1-4

2. Ex Machina (2014) - in class

Week 22 

General Introduction to the Field’s Ethical Questions

1. Sarah Roberts, “Your AI is a Human”

2. Michael Schrage, “Philosophy Eats AI”

3. Molly Smith, “Can Generative AI Chatbots Emulate Human Connection? A Relationship Science Perspective”

4. Meghan O’Gieblyn, “Do We Have Minds of Our Own?”

5. Kyle Chayka, “That New Hit Song on Spotify? It Was Made by A.I.”

6. John Cassidy, “The Dangerous Paradox of A.I. Abundance”

7. Harvard, “How Close Are You to the Way ChatGPT Thinks?”

8. Robert Capps, “AI Might Take Your Job. Here are 22 New Ones It Could Give You”

9. Sam Kriss, “Why Does A.I. Write Like … That?”

10. Bill Wasik, “A.I. Is Poised to Rewrite History. Literally.”

11. Cindy Shan, “Does new 'Cognify' tech allow prisoners to complete years of social rehabilitation in minutes?”

Optional:

1. Ajay Agrawal, “Genius on Demand”

2. Olivia Guest et al., “Towards Critical Artificial Intelligence Literacies”

Independent Reading Project Selections

1. Karen Hao, Empire of AI: Dreams and Nightmares in Sam Altman's OpenAI

2. Yuval Noah Harari, Nexus: A Brief History of Information Networks from the Stone Age to AI

3. Kashmir Hill, Your Face Belongs to Us: A Secretive Startup's Quest to End Privacy as We Know It

4. Ethan Mollick, Co-Intelligence: Living and Working with AI

5. Michael Pollan, A World Appears: A Journey into Consciousness

6. Yanis Varoufakis, Technofeudalism: What Killed Capitalism

7. Shoshana Zuboff, The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power