The Artificial Intelligence (AI) hype train will tell you that we have to adjust to the climate that is the evolution of AI. They will tell you that we have to learn to make it smarter, learn how to use it, and by doing so, we will drive innovation and become an overall more efficient society. But at the expense of what? We are losing the creative process of thinking, speaking, and creating. We forget how to read and analyze. We don’t know how to slow down and trust our own thoughts, ideas, and opinions. We don’t trust others. We don’t trust ourselves. 

How Does AI Actually Work?

To better understand why it has the hold on us the way that it does, it’s important to understand how it actually works and the history of AI, and in this study, generative AI (GenAI). GenAI is a “machine-learning model that is trained to create new data,” one that “learns to generate more objects that look like the data it was trained on” (Zewe, 2023). This means that AI takes in tremendous amounts of information, whether that be through program training or user training. It uses this data to learn the patterns, using those patterns to produce content that resembles the input—sometimes, even, becoming trained to sound as though that original input is yours. The more data that it’s trained on, the more accurate—and even convincing— it may be. 

Why is it so Successful?

In simple terms, convenience. In a world that moves so tremendously fast, the appearance of something will stick more than the depth of it. As a society, we have become so familiar with efficiency that we have a need for speed; we have to go so fast that we have no idea how to actually slow down. An efficient way to analyze this is by studying attention economy, which is the system in which our attention is considered a commodity, “captured, analyzed, and traded for profit” (Masoud, 2025). Our attention requires time, and time is money, and it is quite the hot commodity. We are trained to pay attention to what’s fast and easily digestible, which aligns directly with our society’s demands. 

Our Lack of Comprehension Skills and Media Literacy 

According to the United Nations Children’s Fund, a literacy rate is defined by the percentage of the population that can both read and write a short, simple statement. When comparing literacy rates in youth aged 15-24 between 2018 through 2020 in 202 countries and territories, 99 of those are considered “less developed,” 47 are considered “least developed,” whereas 6 are not classified and only 50 considered “more developed” (UNICEF). While there has been a relative increase in literacy rates since the mid 1900’s (Roser, n.d.), it’s important to ask ourselves why despite this increase in technology, there is still a lack of literacy across the globe. 

In a more recent report completed by The Organisation for Economic Co-operation and Development (OECD), they conducted a survey in literacy in adults, comparing the mean scores of literacy and numeracy in 2011/2012 compared to 2022/2023. Tej Parikh included a graph of this below in his opinion article, “The human mind is in a recession,” for the Financial Times.

(Parikh, 2025)

Just because there are signs of plateauing, does not mean this isn’t a sign of concern. We see decreases. We see the plateau. But where is the increase? Despite an infinite amount of resources that are supposed to make us more well versed, we don’t seem to be getting any smarter. There are countless exogenous factors— a lack of education and access, the political climate of curriculum, and mental health access, to name a few—but one theory that is important to recognize is the use of our muscles. When you stop using them, they get weaker. This is happening in our brains. 

Take college students, for instance. Some universities allow it: students can disclose their use of AI (Felicity, 2024), but others originally completely banned it over fears of plagiarism (Wood, 2023). One of the most frequently cited papers analyzing this phenomenon is an article written by The Atlantic staff writer Rose Horowitch, “The Elite College Students Who Can’t Read Books.” In Horowitch’s article, she discusses how different universities are struggling with getting their students to read.

Nicholas Dames, professor at Columbia University who has taught Literature Humanities since 1988 has seen students become overwhelmed, unable to complete the assigned readings throughout the semester without protest; Twenty years ago, this would not have been a problem. Horowitch writes, “no comprehensive data exist on this trend,” “but the majority of the 33 professors I spoke with relayed similar experiences” (Horowitch, 2024). 

Horowitch continues to highlight the experiences of university professors across the globe: Professor Anthony Grafton of Princeton noticed his students with a weaker vocabulary, Professor Jack Chen of Virginia University struggles as his students “shut down” when confronted with challenging ideas… and the list goes on. 

It’s difficult to study this trend as there is simply not enough data. But what we can notice is that due to educational curriculum and lack thereof, students are simply not entering higher education with the skills needed to complete collegiate level courses. We can quibble about the details, but it is proven over and over that whether it is with or without AI, literacy rates are dropping at a sharp decline, especially in the western hemisphere.

The Decline of Skills

 In a 2023 interview with journalist Rebecca Jarvis, Jarvis asks OpenAI—creator of ChatGPT— CEO Sam Altman about the use of AI; rather than a replacement for human intellect and creativity, Altman described his creation as “an amplifier of the human will” (Mazilu & Popa, 2023).  But yet, it threatens to have the opposite effect. 

Deskilling, as described by Microsoft partner Danah Boyd, is the process of losing specialized skills as well as basic problem solving skills. In creative and technical professions alike, GenAI is shifting society’s relationship with work from participatory and stimulating to passive. 

We live in a time that feels so urgent that it feels like we don’t have time to sit down and learn a skill. Write a paper. Create a painting. When we have to go through the process of learning a skill and we have to face failure, we get frustrated with ourselves and can’t get ourselves to go through it because we crave low effort solutions and instant results. But the process of learning a skill requires time, patience, and above all, failure. We have to sit with what makes us uncomfortable and make mistakes; however, rather than doing so, we prompt a machine to get the results we want in seconds, leading us down the path of “deskilling.”

In Support of AI.. 

Despite the growing concerns of the role of AI in higher education, there are many arguments that it can be useful when integrated into learning environments. Altman states that the ChatGPT should be considered "an incredible educator in our pocket,”  "a new way to supplement learning", and promise "to provide great individual learning" (Mazilu & Popa, 2023). Many journalists, writers, and educators alike argue that Altman’s thoughts of AI should be what it truly is. 

The most consistently argued case for AI usage is its convenience. For students who are struggling in class or with personal responsibilities, they can quickly prompt a chat bot to summarize readings or provide quick explanations. Another frequently cited use of AI is its ability to frame and reframe topics, especially in writing. 

In an opinion blog by writer Emma Dhesi, she argues that “AI is beneficial to writers,” and when she “can take a plotline I’ve got, feed it into AI and ask, ‘What do you think? Where are the plot holes in this? Where can you see things that don’t match up in this,” she can sometimes use AI to spot things that she may not have otherwise noticed (Dhesi, n.d.). She continues: “...even if you choose to write your book using entirely AI, you’re still the creative director. It’s still your name on the page and it’s still your ideas that are coming to the forefront” (Dhesi, n.d.). Rather than replacing originality, Dhesi argues that it has the potential to amplify it. 

Furthermore, employers are looking for individuals who know how to prompt and use AI; It is increasingly present and accepted in the professional environment. In a survey done by Washington State University Carson College of Business and Research—in partnership with KRC Research—researchers explored what professionals were looking at toward AI understanding and usage and found that 88% think U.S. colleges and universities should provide educational opportunities for students to learn about AI, and another 74% believe graduates should already have experience with AI before entering the workforce (Employers Want New Grads With AI Experience, Knowledge, 2024). Employers want students who can use it responsibly, knowing how to prompt, evaluate, and refine AI outputs. Technically, it has also “unquestionably revolutionized our ability to process large datasets and classify or otherwise learn from their content” (Gonsalves et al., 2023).

Regarding the growing concern of the AI misinformation panic, a study done in the Misinformation Review at Harvard Kennedy School argues that while there are real concerns, the growing fear about GenAI’s impact is overblown. It’s argued that there is already misinformation online, but that despite it being so easy to access, “the average internet user consumers very little of it” (Acerbi et al., 2022), and whether users are consuming it or not, “it only gains casual powers” when they see it, yet “the number of things that go on the internet and get seen is finite because our attention is finite” (Jungherr & Schroeder, 2021). They argue that there is always a panic with the introduction of new technologies, but we mostly consume content from mainstream media sources, such as TikTok and Instagram, rather than basing our knowledge and opinions on an AI bot. 

Exaggerated Rhetoric May not be so far From the Truth

While many researchers, experts, and tech leaders argue that the dangers of AI are overblown, real world developments continue to suggest otherwise. What continues to be dismissed as alarming, exaggerated rhetoric now appears to almost ground itself to reality. AI is not only shaping how we think and work, but whether we even work at all. 

As mentioned previously, AI’s continued success is due to its efficiency, but Boyd counters the argument of efficiency, arguing that “efficiency isn’t simply about maximizing throughput. It’s about finding the optimum balance between quality and quantity” (Boyd, 2023). However, in education, this balance is rapidly deteriorating, threatening to disappear before we even recognize it’s happening. Yes, we can argue that our students are gaining speed, but at the cost of critical engagement and comprehension. 

In analyzing the role of artificial intelligence in university, Dartmouth College categorized key words in understanding AI, one of which is known as hallucinations, or “generated outputs that are presented or interpreted as fact, but are in fact merely statistically plausible and inaccurate or incorrect”  (Keywords for AI Literacy | Writing, n.d.). Nowhere is the decline of skills more evident than in education. When students use AI and don’t question what they are receiving, they threaten to lose their critical thinking skills and believe these hallucinations they are receiving. 

The arguments in support of AI seem to wane in comparison to the truth. Altman himself mentioned that “the elimination of jobs, increased racial bias, large scale disinformation, and offensive cyberattacks” are all possible downsides of AI (Mazilu & Popa, 2023).

Consider the labor market, where AI’s impact is no longer speculative, overblown rhetoric. In a 2024 article from Forbes, they stated that nearly 65,000 job cuts were announced, likely due to AI (Roeloffs, 2024). Research done by Stanford  University just a few days ago reports that “employment for workers ages 22 to 25 in the most AI-exposed jobs, including software development and customer service, has dropped by 13% relative to less-exposed roles” (Vaziri, 2025). The threat is prevalent, and very, very real. 

On the Environment? 

The United Nations Environment Programme (UNEP) listed ways that the use of AI is problematic for the environment, including data centers. Data centers take a heavy toll on the planet, making a “2kg computer requires 800kg of raw materials.” These said data centers produce electronic waste, “which often contains hazardous substances like mercury and lead” (AI Has an Environmental Problem. Here's What the World Can Do About That., 2024).

AI uses so much water that it may be more than one consumes in their lifetime. It is assumed that AI is estimated to consume six times more water than Denmark, according to research done at Cornell University. (Li et al., 2025). This is quite the problem, especially when a quarter of humanity lacks access to clean water and basic sanitation. (WHO & UNICEF, 2023). While AI can deliver benefits, the environmental cost is damaging. 

“AI’s water footprint can no longer stay under the radar and must be addressed as a priority as part of the collective efforts to combat global water challenges” (Li et al., 2025). 

 A Deeper Analysis of the Role of Generative AI: Deepfakes

One of the first threats of GenAI was its ability to manipulate and create an image. Even as far back as 2004, journalists and technologists issued early warnings about the role AI plays in the visual world. “It doesn’t take much skill to make the unreal seam real,” Katie Hafner wrote in a New York Times article. (Hafner, 2004). 

A deepfake, as defined by the National Security Agency of the United States (NSA) and the Central Security Service (CSS), is a form of media that has been falsely created or manipulated using some form of artificial intelligence. (NSA, U.S. Federal Agencies Advise on Deepfake Threats, 2023). The most harrowing characteristic of deepfake technology is its disproportionate targeting of women. More than 90% of victims of deepfakes are woman, and because of that, they are “subject to online abuse and sexual harassment resulting from nonconsensual deepfake porn” (“The Dangers That Deepfakes Pose Towards Women,” 2021). Furthermore, in a report done by Deeptrace, it was found that out of 14,678 deepfake videos, 96% were nonconsensual intimate content and that 100% of the “top five deepfake pornography websites” targets women (Ajder et al., 2019). 

This consistent use of objectifying and commodifying women without their consent is turning our bodies into malleable assets for exploitation. This is especially evident when evaluating the use of public figures (such as Taylor Swift, Scarlett Johansson, Billie Eilish) on deepfakes, threatening their reputations and mental health. One journalist, Rana Ayyub, had her face inserted on a pornographic video, and in turn, she was flooded with threats, people propositioning her, and urging others to be sexually intimate—nonconsensually—with her (Giseke, 2020). This is more than an individual harm issue. This does more than damage mental health and reputation. It threatens consent, identity, and privacy. This threatens to perpetuate violence against women ten-fold. 

How can we Combat This? 

As Frank Herbert, the best selling science-fiction author of Dune warned, “Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them” (Herbert, 1999). I challenge you to stop the train and tear down the tracks. I challenge you to be the driver of the train and lay down new tracks, tracks that encourage education, critical thinking, and connection. Tracks that empower women and thinkers and anyone victim to any kind of generated violence that should not have taken place to begin this. We can combat this by becoming disgustingly educated, by working our brain to do the things we once loved doing:

• READ.: The American Time Use Survey shows that “people who read books for pleasure has shrunk over the past two decades” (Horowitch, 2024), despite the rise in book sales (Armstrong, 2022). We have the time and consciousness to know we want to read, to know that we want to invest in these hobbies, but we don’t. We take the easy way out with doomscrolling and generating responses to short prompts in our classes. We skim through a 6 minute article or sparknote our class reading assignment. Keep that reading streak. Read what challenges you. Read something that you stay away from or something that you haven’t read in years. 

• DO WHAT YOU LOVE: When was the last time you really invested in your hobby? No distractions, no fear. Even if it’s a new hobby you want to learn, allow yourself to make mistakes. Fail. Take back your focus. 

• TRANSPARENCY: Not only do we need to be critical thinking and engaging with the news and media we consume, but we need to be transparent in our dissemination of knowledge and reporting. I know you know someone who uses AI like it’s their therapist, their tutor, or an extension of themselves. Inform them about the repercussions of this and allow them to become better learners and thinkers because of it. There is, and always will be, mixed opinions about the use of AI, but having authentic data and reporting on it will inspire a trend of transparency. Promote intentional use of AI rather than default use. 

• REINVEST IN HUMAN CONNECTION: Not everyone has access to mental health resources, books, art, and tools in general they need to be happy and healthy humans. But rather than gutting these things, we need to re-invest in human connection. Talk to your teachers. Advocate for better educational systems. Work with your friends when you’re stumped rather than asking your question to ChatGPT. Pay artists to create. Make space for learning, failure, reflection, and connection. 

Artificial Intelligence can do many things, but it cannot care. It cannot dream. It cannot connect. That is our greatest strength.