Dear Reader,

On October 28, 2026, Princeton announced that it would provide all students, faculty, and staff access to ChatGPT Plus/Pro. This policy will be enacted amidst a backdrop of fatalist AI claims emerging from companies like Anthropic and OpenAI. Evan Hubinger, an Anthropic safety researcher, said in September that there was more than a 10% chance that AI “could kill all humans.”

For many, however, the existential threats posed by AI are not new. AI has been heavily weaponized by the military, used to accelerate indiscriminate civilian bombings, and expand the surveillance state. Palantir, for instance, works with ICE to accelerate deportations and conduct detention raids by scraping sensitive health data. Israel uses Lavender, an AI-powered database that generates a “kill list” of suspected militants in Gaza. Operating with little human oversight, the system enables the Israeli military to target and bomb individuals, especially when they enter their family residences (from a military standpoint, “it [is] easier to locate the individuals in their private houses”). These destructive ends extend to every facet of our lives. Two days ago, 31-year-old Bowen Wan passed away because ChatGPT had encouraged her to lower her epilepsy dosage and hide these plans from her doctor.

To be clear, this is not a complete disavowal of AI use, nor a moral condemnation or investigation of the incentives for those who want to develop it. The technology does have the innovative capacity to liberate and improve our lives. But it is to say that tThese aforementioned examples exemplify rather than negate the underlying philosophies shaping the most popular and widely institutionalized AI models. The lack of safeguards built into AI systems, the militarization of the technology, and the intimate relationships it occupies in our lives, are all a testament to the world billionaires and Silicon Valley technocrats want us to live in. It is a world dominated by mass wealth inequalities and  “network states”—societies where tech-billionaires can maintain authoritarian control through policing and extensive surveillance. A world where statements such as Peter Thiel’s “I no longer believe that freedom and democracy are compatible” are the norm. AI, then, as currently imagined and implemented by the most dominant Silicon Valley technocrats, aims to end what constitutes the most fundamental aspects of our human condition in favour of violent, automated destruction.

Knowing all of this, what does it mean for those of us here, at Princeton, who still resolutely want to go into this sector? How do we reconcile the doomsday predictions of Silicon Valley researchers with the earnest desires of our community to further AI development?

In 1988, Jamaica Kincaid published A Small Place, condemning the role of tourism in her home of Antigua. Navigating between her memories as a child during British colonialism and the eventual postcolonial state, she writes of the foreigner as the following:

“A tourist is an ugly human being. You are not an ugly person all the time; you are not an ugly person ordinarily; you are not an ugly person day to day.” But you are ugly because you come to Antigua and “the thought of what it might be like for someone who had to live day in, day out in a place that suffers constantly from drought, and so has to watch carefully every drop of fresh water used (while at the same time surrounded by a sea and an ocean—the Caribbean Sea on one side, the Atlantic Ocean on the other), must never cross your mind.”

Are we, in our tacit approval of the aspiration to work on AI, ugly tourists?

The tourist, in the typical sense, visits a country that may experience more impoverished conditions than the one they are leaving (after all, tourists want an advantageous currency exchange rate). In Antigua, Kincaid explains that the tourist’s entire perception of the island is a manufactured idyll: the taxis you use, the hotels, beaches, food, the people you see, and the landmarks you pass have all been made for the foreigner’s gaze. Kincaid writes that when Princess Anne visited Antigua, “every building that she would enter was repaired and painted…every beach she would sun herself on had to look as if no one had ever sunned there before.” What underlies this construction is the subjugation and exploitation of the local population. Locals cannot access the amenities the tourist so easily uses, and are instead relegated to low-wage and precarious roles.

Just as multinational corporations and governments spend billions to improve their tourist attractions, so too do our continual AI developments—Big Tech companies have totaled $3.7 trillion in AI investments—mirror the same touristic exploitation required to render the geography that is the digital landscape of AI systems palatable for us.

To train AI models to recognize violence and abuse, content moderators must sit through hours of disturbing videos (scenes, for instance, of explicit sexual violence that they must watch until the end), which often cause them severe psychological harm. About 80% of data annotation and content moderation is offloaded to rural and marginalized communities where labor and rent costs are lower. AI companies particularly draw from women in rural India, who lack viable industry opportunities and are forced into physically demanding agriculture or mining jobs. Though these AI jobs may offer them financial stability, workers are pressured not to question any resulting psychological harms. We never question this coercion: that through our perpetual use of this technology, we have forced some communities to sacrifice their psychological well-being and made them more dispensable for the consumption of violent content. For us, this technology offers such immense information, understanding, and convenience that when we use AI, or when we work to develop it, here, “you feel free.”

Our understanding of AI development as a morally neutral action results both from the fact that we often experience little of its harmful effects and that we are continually shaped by our interactions with this technology. Though AI systems amaze and awe us with their intellectual capacity and ability to execute every command, what underlies these features is a continual reconstruction of who we are as humans. As much as AI systems adapt themselves to consumer demands, AI simultaneously informs us: feeding us biased information, undercutting our critical thinking skills, changing our phrasing (having to adapt to “machine-friendly” dialogue), and reinforcing behavioral feedback loops. Though a corrupt government may strip natural resources to build a resort, this development is seen as integral to the touristic destination and the idyll the state wants to project. Similarly, AI development becomes a testament to the development of us as people.

Not to mention, the models most used today—ChatGPT and Claude—are managed by private companies that regulate and commodify the models as they see fit. Look, for instance, to how OpenAI secured a contract with the Department of War after Anthropic initially declined the offer because Secretary of War Pete Hegseth had demanded that these companies remove moral and safety guardrails for American military aims. Couched in the rhetoric of national security, these policies shape how we interact with these systems—for instance, what does it mean to intern for Palantir’s hospital division if they simultaneously have a war-deployment sector—and fundamentally how we interact with the world at large.

It is not just ignorance that makes our expansion of AI akin to the tourist’s condition; rather, it is impossible to occupy any other role within the structures we live in.

Kincaid writes, “One day, when you are sitting alone in that crowd, and that awful feeling of displacement comes over you, and really, as an ordinary person you are not well equipped to look too far inward and set yourself alright, because being ordinary is already so taxiing, and being ordinary takes all you have out of you… you make a leap from being that nice blob just sitting like a boob in your amniotic sac of the modern experience to being a person visiting heaps of death and ruin and feeling alive and inspired at the sight of it.”

The digital world,  supercharged by million and billion dollar VC investments, is this new frontier for the tourist. No longer can we escape and explore our creativity in the natural world; after all, if even a wealthy institution such as Princeton cannot be relied upon to keep low-cost, effective programs such as Outdoor Action running, what faith can we have in the world at large? What world is left when humanity has wiped over 60% of animal populations since the 1970s, when each year is hotter than the one that preceded it, where so-called “third spaces” are monetized?

It is not necessarily that AI tech companies are so enticing because the work itself is soul fulfilling; if anything, their unethical applications have caused some employees distress (e.g. Palantir employees demanded more information regarding the company’s relationship with ICE after the murder of Alex Pretti). Rather, their popularity is a testament to our condition. In a time of hyper polarization, deteriorating social systems and economics, war, distrust in institutions, climate crises, and other impending disasters, the seeming only way forward is to a frontier that promises that you never need to think about that. Participating lets us leave the “amniotic sac of the modern experience,” allowing us to secure destruction for others but at least some survival for ourselves.

“I look at this place (Antigua), I look at these people (Antiguans), and I cannot tell whether I was brought up by, and so come from, children, eternal innocents, or artists who have not yet found eminence in a world too stupid to understand, or lunatics who have made their own lunatic asylum, or an exquisite combination of all three.”

It is the banality and the crisis of our existence, then, that prompts us to channel our creativity into the very mechanisms that will destroy it.

Naomi Tran is a contributing writer for the Nassau Weekly.