UC Berkeley campus transformed into a cesspool of pro-AI lobbyists when the academic year began. Simply walking down Sproul Plaza, one is surrounded on all sides by swarms of registered student organizations and internship pushers promoting “AI-optimization coding” or “AI business consulting.” The university continues to open more buildings such as “The New Gateway,” described as a “dedicated UC Berkeley hub for AI.” I often hear peers discuss their use of AI for various tasks, from summarizing readings to generating entire essays. It certainly seems that AI is ubiquitous across workplaces and classrooms in UC Berkeley.

Nonetheless, this large language model trend is not without its critics. Often, when I’m alone with my artsy, leftist, tote-bag-toting friends, we will indulge in long conversations about the moral and ethical implications of AI between sips of iced coffees or puffs of a prerolled joint.

Why, one must ask, do American students and professionals prioritize the ease provided by LLMs above the well-being of the health of our air and water and the lives of the workers bearing the brunt of AI operations under the specter of colonization?

But my AI-abstinence is a minority practice among my peers, with 57% of college students in the United States using the technology weekly. And it’s not entirely their fault.

Realistically, if you want to go into consulting, marketing or tech fields, your boss will require that you use AI, whether you would like to or not. I can see how hopeless the fight against AI may seem, especially when the UC system is materially invested in it, allotting $18 million to “scale up” AI last May and declaring AI advancement as part of its “long-term investment strategy.”

In the bleak shadow of Silicon Valley tech tycoons, I was happy to see my professors addressing AI use in class, and even posters advocating that students “QuitGPT.”

Last semester, my professor awkwardly questioned students about AI. At first, the class was silent. It felt like I was back in middle school, being taught sex ed for the first time, with neither the teacher nor students particularly keen to discuss the topic or their experience with it.

Still, I assumed that, even if most students used AI, at least our history-major writing classes would be sheltered from this vice.

I was wrong.

After a few brave souls raised their hands, the professor asked us to pull up Google Gemini on our UC Berkeley emails and ask it about his new book. He wanted to see whether it would provide an accurate summary of the argument and themes. I shot a hesitant look at my friend beside me. To overextend a metaphor, I pride myself on my Regina George-esque half-virginity when it comes to using AI, having only engaged with it through backdoor options, like Google’s inescapable AI overview. Now my virtue was being threatened.

The rest of the class time was spent discussing the “safe” and “healthy” ways to use AI. It made me question — was my abstinence-only approach to LLMs doing more harm than good?

Simultaneously, I started seeing more QuitGPT propaganda around campus and on my social media. According to its website, QuitGPT’s goal is “organizing Americans and people around the world to quit ChatGPT.” It emphasizes that Sam Altman, the CEO of OpenAI, has contracted with the Pentagon and supplied it with military and surveillance technology to maximize his company’s profits.

I am not one to defend OpenAI, the Pentagon or Sam Altman. But QuitGPT is not anti-AI. In a pinned post on its Instagram account, it encourages its followers to switch to Claude or Gemini instead of Altman’s alternative, explaining that “Claude and Gemini aren’t innocent, but they’re not Trump megadonors.”

What this logic fails to consider is the equally negative impacts of ChatGPT’s main competitors. In addition to the devastation posed by all AI companies and data centers — namely, rising utility costs and environmental ruin to nearby communities — Google Gemini recently invested 40 billion in Anthropic, the company responsible for Claude. Anthropic has been under scrutiny for contracting with the Pentagon in July 2025 to provide the United States government with surveillance and “autonomous weapons” technology, even being used by the U.S. to aid the capture of Venezuelan President Nicolás Maduro this January.

In fact, Anthropic’s main dispute with the Trump administration stemmed from its AI system being too violently unbridled. According to an April report from the Pentagon, the autonomous weapons systems provided by Anthropic stray from the U.S. Department of Defense Directive, or DoDD, by not “allow(ing) commanders and operators to exercise appropriate levels of human judgment over the use of force.” The DoDD claims that this constitutes “a supply chain risk to America’s national security.” In response, Anthropic challenged the DoDD in court. The legal proceedings are still ongoing, and due to a court-issued preliminary injunction, Anthropic and Claude are still actively working with the Pentagon.

Claude and Gemini, now materially intertwined through billion-dollar investments, are much worse than “not innocent.” They are destructive, murderous and in cahoots with the United States’ fascist government. And despite QuitGPT’s claim that their mission is to make all CEOs “think twice before they get in bed with Trump” by targeting Altman and OpenAI, in reality, they will likely only be a boon to ChatGPT’s competitors who are already cozy with the Pentagon.

At the end of the day, AI is not like sex. It is safe and practical to advocate abstinence.

Although there can and should be nuance in the discussion of individual LLM use, especially when required by a class or a workplace, there is no place in the anti-AI movement for a half-hearted boycott. Just as it makes no difference to choose whose gun will shoot you, it is not productive to advocate for Gemini and Claude over ChatGPT or other AI systems. To resist AI is a radical act, and we must have radical demands.