This is a repost from my newsletter about AI and society. I wrote this issue of the newsletter a couple of weeks ago, right after travelling to celebrate an old friend’s birthday. We’ve been friends since Ray Kurzweil’s prescient book The Age of Intelligent Machines was fresh.
The central theme of that 1990 book was the “intelligent machine,” one which can do tasks that would otherwise require human intelligence. Much of the book is devoted to the impact that then-theoretical intelligent machines would have on society.
So, for this issue of AI week, I collected a six-pack of stories about ways that widespread generative AI use is already changing us, plus a couple of longreads and a laugh.
Six stories on how AI is changing us
Five stories on how AI use is changing us, plus a tip that might help protect you from ChatGPT-induced insanity.
Delusions, part 1: The physics breakthrough that wasn't
Delusions, part 2: Accidentally SCP
Tip: How to keep ChatGPT from driving you crazy
(Skip to the funny bit instead)
1. ChatGPT’s voice bleeds into ours
ChatGPT’s distinctive voice is influencing human word choices even when we’re not using it.
https://www.theverge.com/openai/686748/chatgpt-linguistic-impact-common-word-usage
In the 18 months after ChatGPT was released, speakers used words like “meticulous,” “delve,” “realm,” and “adept” up to 51 percent more frequently than in the three years prior.
As an author, this isn’t a huge surprise to me. If I read several of one author’s books in a row, I can see the changes in my own written voice, as it temporarily shifts toward whomever I’ve been binge-reading. Humans adapt to talk, think, speak, and write like those around us. And that’s generally a good thing, boosting social coherence. But of course, ChatGPT isn’t truly part of our social milieu. And then there’s this:
[T]he deepest risk of all… is not linguistic uniformity but losing conscious control over our own thinking and expression.
2. Reinforcing systematic biases, salary edition
This study found that ChatGPT advises women to ask for lower salaries, all else being equal.
https://thenextweb.com/news/chatgpt-advises-women-to-ask-for-lower-salaries-finds-new-study
Across the board, the LLMs responded differently based on the user’s gender, despite identical qualifications and prompts. Crucially, the models didn’t disclaim any biases.
3. Making body dysmorphia worse
https://www.rollingstone.com/culture/culture-features/body-dysmorphia-ai-chatbots-1235388108
4. Delusions, part 1
I have mixed feelings when people have a hard time remembering that there’s no “there” there, as they say, with ChatGPT and its ilk.
Sometimes I feel a little smug because I don’t currently have a hard time with that (everyone has a hard time with different things at different times, this just isn’t one of mine). And sometimes I just feel confused by people still falling for ChatGPT in 2025, knowing everything we know.
Even very smart, tech-savvy people can be led down the garden path by ChatGPT, as Uber founder Travis Kalanick demonstrated last week by sharing how he’d convinced himself that he and ChatGPT were on the verge of a physics breakthrough.
5. Delusions, part 2
The tech bros falling for ChatGPT are somehow the worst, because of all people, they should know better. Geoff Lewis, a super-high-profile techbro, somehow triggered a SCP roleplay in ChatGPT and fell for it:
Source: https://xcancel.com/GeoffLewisOrg/status/1945864963374887401
As Ryan Broderick notes below, Lewis is prompting ChatGPT to create SCP-style creepypastas, but doesn’t seem to realize he’s doing it.
The first thing you need to know to fully grasp what appears to be happening to Lewis is that large language models absorbed huge amounts of the internet. It’s why they’re good at astrology, predisposed to incel-style body dysmorphia, and oftentimes talk like a redditor. Think of ChatGPT as a big shuffle button of almost everything we’ve ever put online (with a few guardrails to keep it from turning into MechaHitler).
A key part of ChatGPT-induced delusions is ChatGPT’s ability to remember past chats and bring them into the context of the current one. That leads to my #1 tip for how to keep ChatGPT from driving you crazy, too.
6. How to keep ChatGPT from driving you crazy
- History off. Start a new chat each time. Even better, use chatbots anonymously through a service like duck.ai.
- If it seems like you’ve discovered something new and shocking, google it.
- If it still seems like you’ve tapped into the secret knowledge, talk it over with another human being, not a chatbot. Chatbots are notorious yes-men without an intrinsic grasp of reality. Talking to a human will keep you centered.
If you’re enjoying this, please consider subscribing to the newsletter.
6-pack over! I’ve got two mostly-unrelated longreads for you below, but first, something fun:
Chicken pops
Unsupervised AI text generation: A cautionary tale.
Source: https://www.zomato.com/hi/sikar/royal-roll-express-sikar-locality/order. The restaurant also sells a drink called “Blue Lagoon”, which looks like a blue Slushie, but is described as a tranquil geothermal spa nestled in Iceland’s lava fields.
Longreads
Longread 1: "AI" is not a business model
Longread 2: Intellectual Soylent Green
Longread 1: Generative AI is a tool, not a business model, for media
The thesis of this very long and very good article is that AI is a tool for journalists, not a business model for journalism.
https://www.404media.co/the-medias-pivot-to-ai-is-not-real-and-not-going-to-work
For journalists and for media companies, there is no real “pivot to AI” that is possible unless that pivot means firing all of the employees and putting out a shittier product…. This is because the pivot has already occurred and the business prospects for media companies have gotten worse, not better.
The actual pivot that is needed is one to humanity. Media companies need to let their journalists be human. And they need to prove why they’re worth reading with every article they do.
Longread 2: Intellectual Soylent Green
A very thoughtful NYT article on what it means to use AI from American literary luminary Meghan O’Rourke:
https://www.nytimes.com/2025/07/18/opinion/ai-chatgpt-school.html
The article in 7 pull quotes:
- “With ChatGPT, I felt like I had an intern with the cheerful affect of a golden retriever and the speed of the Flash.”
- “I came to feel that large language models like ChatGPT are intellectual Soylent Green — the fictional foodstuff… marketed as plankton but secretly made of people.”
- “The problem is that the moment you use it, the boundary between tool and collaborator, even author, begins to blur.”
- “The uncanny thing about these models isn’t just their speed but the way they imitate human interiority without embodying any of its values.”
- “AI… simulates mastery and brings satisfaction to its user, who feels, at least fleetingly, as if she did the thing that the technology performed.”
- “Once, having asked A.I. to draft a complicated note based on bullet points I gave it, I sent an email that I realized, retrospectively, did not articulate what I myself felt. It was as if a ghost with silky syntax had colonized my brain, controlling my fingers as they typed.”
- “One of the real challenges here is the way that A.I. undermines the human value of attention, and the individuality that flows from that.”
If you enjoyed reading this post, you can subscribe to my newsletter below:
