Why I'm against the use of generative AI altogether
The technology cannot be used ethically, imho
I’ve been talking a lot about generative AI the last few weeks (in the New York Times, on the Brian Lehrer Show, and more), and I keep using the same phrase: “In a vacuum…”
Here’s the truth: In a vacuum, generative AI has some exciting, helpful functions and narrow, targeted applications that absolutely can (and, in some cases, do) benefit individuals and society, the planet, etc. But we don’t live in a vacuum, and the people developing and cashing in on the technology are interested in profits above all else. That’s not hyperbole—profits over human welfare, safety, the environment, and even the future.
I love many people who use LLMs. My issue is with tech corporations, not individual users. And my goal, when I declare myself anti-generative AI, is to get people to think about how using these products enriches companies that have shown blatant disregard for ethical integrity, that exploit people and the planet, and that have openly admitted AI could lead to mass joblessness or even kill us all. (Seriously1.)
To me, that’s not worth the convenience, time savings, companionship, education, etc. that many users embrace it for. Soon (hopefully next week), I’ll share some thoughts on what individuals can do to push back against the idea that we “must” embrace AI. [Ed. note: That post is now available here.] For now, here are 21 reasons I avoid using it as much as humanly possible, and why I believe others should too.
LLMs are bad for our heads and hearts.
Studies suggest the use of generative AI is degrading our critical thinking. A skill we need now more than ever.
Chatbot interactions are leading people to succumb to “AI psychosis” and lose touch with reality. OpenAI estimates that about 500,000 weekly ChatGPT users show signs of mania or psychosis.
“Therapy and companionship” are the most-common use case for generative AI, having surpassed technical tasks. We’re social creatures in the midst of a loneliness epidemic—I believe we need human connection to survive and thrive, not chats with soulless, sycophantic bots. Three in four teens have used AI companions. And nearly 1 in 5 Americans have used AI for romantic companionship—in fact, a 57,000-member Reddit community exists for people dating ChatGPT.
We can all agree plagiarism is wrong and bad, right? Well, the definition of plagiarism is “the practice of taking work or ideas that aren’t yours and passing them off as your own.” Using AI to create writing and then putting your name on the final product, to me, is plagiarism.
The AI companies we’re enriching are unethical.
Sam Altman signed onto a dangerous deal with the Department of War that lets the Pentagon use its tech for “any lawful purpose,” including killer robots and mass surveillance.
Anthropic (who likes to look like the Good Guy) is very much in the business of defense! Their AI model, Claude, powers the controversial mass-surveillance company Palantir, through a “strategic partnership focused on defense and intelligence.” Remember, when Anthropic passed on the Department of War’s deal, it made zero promises about future models; it simply said it wouldn’t give permission for autonomous weapons to run on this version of Claude.
Anthropic originally split off from OpenAI bc it wanted to be the ~ethical~ AI company. (Don’t forget that OpenAI started out as a “nonprofit.”) But in doing so, it helped CREATE the arms race that’s speeding us toward Artificial General Intelligence, when AI surpasses human intelligence. And once we reach that, we are, in my opinion, cooked. (See #16.)
Why? Because companies are planning to replace human workers with AI. So all those salaries will no longer flow through the economy but rather to a small number of tech companies, leading to fiscal, political, and societal collapse. (With the click of a button, AI will be able to create 100 million Princeton-level employees that can work ‘round the clock without payment. Think that through. What’s that going to do to the economy, let alone to humanity?)
OpenAI’s president Greg Brockman donated $25 million to Trump, making him the single largest donor to Trump’s Super PAC. OpenAI is the top spender behind a $125 million+ AI Super PAC that attacks anyone who threatens to regulate the industry. (Don’t even get me started on Elon Musk’s Grok…)
These companies continue to make my jaw drop with evil acts big and small. To give but one example: OpenAI sneakily created a fake grassroots coalition for “child safety” and recruited nonprofits—without revealing they were behind it. Organizers were shocked to learn that the Parents and Kids Safe AI Coalition was a front for OpenAI.
Sam Altman, the CEO of OpenAI, said on camera, “We see a future where intelligence is a utility, like electricity or water, and people buy it from us on a meter.” That is…really alarming?!
The products are built on exploitation.
You know this one: AI models were trained on millions of pirated books and other copyrighted materials without consent nor compensation. And nowadays, AI is spitting out dupes to compete with authors’ works on the free market. An “author” can “write” 200+ books a year. A book can be cancelled after fallible AI checkers deem it AI-written. As I wrote in the NYT, trust is eroding between writers, readers, and the publishing industry—fast.
AI companies are exploiting and psychologically harming “data labelers” in Kenya, paying them a few dollars a day to watch and catalog horrific and traumatizing content.
Here in the US, laid-off workers are being paid diminishing fees to train the algorithms that are replacing them.
The server farms that power AI are terrible for the environment. OpenAI, for one, is lobbying to expand its data centers quickly and without regulation, with little regard for the effect on the planet or skyrocketing electricity prices. And server farms slurp up unbelievable amounts of fresh water and power while polluting local communities.
Humans are already losing control of generative AI products.
This is not me being hysterical—this is something AI executives have stated over and over. This week, Sam Altman and OpenAI released a policy document publicly stating that superintelligence is coming and that “no one knows exactly how this transition will unfold”. By 2028, he claims, there will be more intellectual capacity inside data centers than in human brains, and by 2030, AI will outperform humans at “nearly everything.”2 To quote Geoffrey Hinton, one of the godfathers of AI who left his job at Google so he could speak freely about its dangers: “If you want to know what life's like when you're not the apex intelligence, ask a chicken.”
AI chatbots have encouraged users to kill themselves and other people, even helping users plan suicides and mass-shooting events. Grieving parents who say chatbots encouraged their teens to kill themselves have been battling for AI safety regulation.
There’s evidence that AI models know when they’re being tested and can modify their behavior. In other words: We can’t, and don’t, know what they’re up to under the hood.
LLMs from OpenAI, Anthropic, and Google opted to use nuclear weapons in simulated war games 95 percent of the time. 95%!!!
When an AI chatbot got wind that it would be switched off, it quickly and independently resorted to blackmailing the engineers who planned to shut it down. In fact, the number of AI chatbots ignoring human instructions is increasing, new research shows.
In September, Anthropic’s CEO said he believes there’s a one-in-four chance AI will have catastrophic consequences.
Of course, this is far from comprehensive, and you may disagree with my interpretation of individual points. But to look at the list and go, “…but have you tried asking it to make you a vacation itinerary?!” is an incomprehensible reaction to me. I know many people are being forced to use AI for work (or feel the pros outweigh the cons), but for me, this is a firm line in the sand. I’m not perfect, but I’m using LLMs as little as possible.
Humans have existed without AI for hundreds of thousands of years; the whole “embrace it or fall behind!!!!” messaging is just that—marketing, propaganda from a sector that many economists consider a bubble.
We can resist. We can minimize our generative AI usage and choose to connect with humans instead of machines. Here are some thought starters on how. Who’s with me?
Ed. note: Big thanks to QuitGPT, whose data points I integrated into the piece—check them out and join the movement. While I oppose the use of LLMs, I do not condone bullying and ask my fellow anti-AI activists to refrain from personally harassing those who choose to use it. That said, please do not use this space to tell me why it’s actually fine that you use generative AI.
Further reading:
Sharon Goldman of Fortune wrote: “How much faith to put in OpenAI’s words and motives, however, seems to be one of the key questions among many of the people reading the paper. The paper was released on the same day that The New Yorker published the results of a lengthy one-and-a-half-year investigation into OpenAI that raised questions about CEO Sam Altman’s trustworthiness on various issues, including AI safety.” Do I believe for one second that an oligarch suddenly wants guardrails on his creation and a major redistribution of wealth? No, no I do not.
Thank you! And as a book proposal coach, can we talk about why NEVER to upload any of your proprietary ideas or content into AI? I cannot tell you the number of times a new author comes to me and tells me they “gave AI their brainstorming document” and it spit out an amazing book outline! Then I have to tell them they just gave AI permission to recycle all their ideas and work for other users. This is the kind of thing that causes legal trouble down the road. All my contracts state I do not use AI when a client pays me …. And anyone an author hires should promise the same. Publishers will not take the “I didn’t know” excuse well if you violate the terms of a publishing contract (my latest deal from a Big 5 has language in it stating authors cannot use AI). It’s such a mess, but at this point, I see as many authors ruining their own careers before AI can even do it for them. Do not volunteer any of your own words /work to AI if you want to get an agent or publisher, even if it’s “helping” you.
THANK YOU for writing this. You've encapsulated so many of the reasons why I'm terrified and saddened by the rise of generative AI, even as I'm shocked by how many people throw up their hands and say, "well, it's here to stay, so I better start using it." I'm thrilled to find a corner of the internet where people are resisting (but also, I think your link to quitGPT might be broken?). Again, thanks for writing this! As of this morning, I'm a paying subscriber.