Sorry this is late, lovelies. I took a post-revisions break, then picked up a rush job (corporate graphic design), then got to—well, details are below in Joyful Things.

It’s hard to be a good person. This is a recurring theme in these Musings because it’s a recurring theme in my life. And every time I come up against that thought, I remember the TV series, The Good Place. And that helps deflate the stress and nasty self-talk that’s built up while I strive to be a good person in a world currently dominated by people who have no interest in being good, but only in feeling good about themselves.

If you haven’t watched the series, then please stop reading and come back to this once you have. Seriously. If you’ve somehow managed to avoid spoilers this long, then I am not gonna be the one who messes that up for you. Because you deserve to experience The Good Place as it spools out in all its zany, heartfelt, heartbreaking comedic glory. Go on…shoo. This is your one gentle and firm caution.

🎵 🎹 exit music plays 🎹 🎵

So. There are a lot of things to draw from The Good Place. The thing I want to focus on is just that overarching theme: it’s so much harder to be a good person in our modern age than it seemed to be in the past. And that difficulty lies in complexity, rather than knowledge or intelligence. Being a moral philosopher like Chidi will not magically untangle the skeins of uncertainty that weave themselves through our modern lives. In fact, I’d argue that grey morality is a feature rather than a bug of the systems we find ourselves forced to work within, systems dominated by a small subset of wealthy people increasingly divorced from the realities of life for the rest of us.

Lovelies, I am once again musing about “artificial intelligence.” I have a lot of thoughts on this, lovelies, so I get it if you can’t read this all in one go, but I do hope you’ll stick with me to the end.

Don’t worry, I haven’t changed my stance. There is so much mis- and disinformation around “AI” though, and I do not want to contribute to it. So lemme recap what I know as a concerned layperson, with the caveat that this is far from exhaustive, esp. since things are moving too quickly for any one person to keep up-to-the-minute:

• The conversation about AI existed long before our current hype-bubble-industry. Mostly, if I’ve read my research accurately, it was concentrated in robotics and machine learning circles. Which makes sense, right? We have scads of scifi media that assume AI as a worldbuilding given, powering starships as well as society. The real-world pursuit of robotic simulacra of humans is at the heart of all those books and films and TV series.

• The term “AI” was co-opted (because of its popularity/familiarity from all that scifi media) by some admittedly very savvy marketing people for what I and others call slopbots (Claude, ChatGPT, Copilot, Gemini, Midjourney, etc. etc.) aka “generative AI” tech. I append “generative” to denote the LLM-based tech that’s unfortunately become ubiquitous, crammed into everything from word processing to search engines to appliances. It’s called “generative” because it supposedly generates something from all the data that a human inputs into it, usually in the form of prompts and ongoing chats.

• The foundation of this “generative” tech is Large Language Models. LLMs are the software trained on millions of books, articles, pieces of art, etc. Training in this context means—and I’ll focus on books here—giving the LLMs massive amounts of words and sentences and paragraphs, as data points for its algorithms to work on. I think of these LLMs as meat grinders. In goes trillions of words, meaningfully placed in very specific order, often over the course of years by a human author; out comes a recombined mass, ground into content without context. All denotation, no connotation.

• At one point early in the process, companies like Meta and OpenAI considered whether to legitimately purchase all the books, articles, etc. they wanted to use in order to train their LLMs. Once they considered the price tag and the effort to gain consent from authors, however, they chose instead to use pirated books sites. They used the “fair use” argument to justify disregarding copyright law; as if wholesale theft of entire books were the same as, say, copy-pasting a quotation or paragraph for an academic essay or news article.
  Pirated books sites were already stealing from creators by offering PDFs and ebook files of books for free, without the consent of authors (or publishers), and without any royalties going to authors. I suspect many people who download pirated books may (speciously) justify it as a blow to corporate-conglomerate publishers. But honestly, the people that book pirating hurts most are authors. Unlike publishers, few of us have hundreds of books to earn royalties from. We may only have 1 to 5, and they aren’t bestsellers, either; especially if we’re marginalized.
  Regardless of how many books any one of us may have published, one or a hundred, pirating our copyrighted works is a dick move. It disrespects our legal rights to our own creative works. I think if a reader can’t afford to buy, then use the public library system.

• I know there are authors who don’t have a problem with “genAI” doing the “writing.” I think some of these authors share the mindset of tech slopagandists, that fiction is just a kind of content, and that being able to churn out “books” fast means making money fast. Whether this actually makes them money faster, I don’t know. At this point, it’s clear that “genAI” slop has accelerated the release of ebooks, turning these “books” into widgets. I’ve said it before and I’mma say it again: this is just the most recent point on the graph of devaluing fiction writing, along with low author royalties, and pricing ebooks at $0.99.
  I know I speak from a position of relative privilege; I don’t, and can’t, rely on my writing income to make ends meet. Still. It deeply frustrates me that writers feel pressured to devalue our creative work, in exchange for scraping by in most cases.

• I also know there are authors who mostly want to be compensated for their books being ground into data points for LLMs. This is the conversation around licensing that professional writers orgs like The Authors Guild (USA) and The Writers Union of Canada are pushing; they want authors to get paid to ‘feed’ training models. I disagree with this approach, too.
  I don’t think the Anthropic settlement, for example, is a meaningful or just consequence for stealing creative work. I think it sends the message that stealing from creators is acceptable for whoever can afford the financial penalties—if there’s plaintiff money to pay lawyers to bring a case to the courts; the onus is on the people victimized. That doesn’t say, Respect creators’ rights. That reads as, Don’t get caught.

• The licensing conversation already concedes the point that “genAI” is acceptable. This is the ‘it’s already here; don’t get left behind’ argument. I vehemently call bunk on that nonsense. I will die on the hill of refusing to use “genAI” for as long and as much as I can. If there’s a workaround on my search, I will use it. If there’s autocorrect, I will turn it off. (If it’s turn-off-able; I know, some aren’t.) It is an absolute pain in the ass to do all this, but I refuse to normalize its utility.

• Because “genAI” as it’s been pushed into consumer products is worthless. Worse than worthless: it’s actively harmful.

• If I were to list the documented harms inflicted by slopbots, this is an incomplete list of content warnings I would need: suicide, child sexual assault material, nonconsensual sexual imagery, sexual assault imagery, sexual violence imagery, misogyny, misogynoir, mental psychosis.
  These content warnings apply to human workers tasked with teaching slopware to recognize/categorize/tag traumatizing material; to the innumerable teens, children, and women victimized by deepfaked pornography; to teenagers urged to die by suicide; to adults sunk into psychosis by reasonable-sounding advice that’s really only regurgitated reinforcement of their worst fears and insecurities.

• If I understand correctly, there are useful applications of “AI” in the scientific and medical research fields, for example, where machine learning is applied to collate, summarize, and analyze huge amounts of raw data, in order to distinguish patterns for human researchers to then investigate further.

• “AI” tools have, naturally, written code in mere fractions of the time a human coder would need. Some have lauded the elimination of the tedious parts of coding. I can’t speak to the quality of that resulting “AI”-written code, simply because I’m not qualified to have an opinion. But I will go out on a limb to argue that coding, like many creative outlets, is as much about the process as it is about the end product. A coder’s expertise is built on their facility with coding, in coming up with approaches to creating specific pieces of code to fit specific, desired results, over a career. What happens when humans can’t understand the code an “AI” has written? How does anyone troubleshoot? While the reply to that seems like more “AI,” this only serves to emphasize the current conundrum for human coders.

• All iterations of “AI” are deleterious to the environment. In addition to hazardous waste and unsafe noise pollution, it requires massive amounts of energy, as well as massive amounts of water. The former is due to dedicated data centres; the latter keeps them cool, and in working order. The so-called “AI” boom has meant a huge increase in the pace and scale of new-build data centres, which has meant increasingly displacing resources from human communities. Most often, though not always, these impacted communities are marginalized economically or racially, or both. (And, really, it shouldn’t matter, but we’ve all witnessed that impacted affluent white communities will get faster widespread media coverage than marginalized ones, regarding the same issue.)

• This technology replicates the baked-in biases of the material it’s been trained on: misogyny, racism, queerphobia, etc. These biases are invisible within the tech itself, having been inserted with its foundational training materials. (Since publishing also skews heavily through the ages toward white men, guess whose biases got siphoned most into LLMs?) This in turn reinforces the harms of systemic inequities, allowing common users and slopagandists alike to pretend such bias doesn’t exist. Which often ends up forcing us to remind them, wasting our time and energy over and over again. (To borrow from and paraphrase Toni Morrison.)

There’s a question I’ve taught my children to ask themselves when they face a difficult task: Is it worth it? Is it worth the sweat, the brain-melting effort, the heartache, the [fill-in-the-blank], to do this thing, whatever it is?

In other words, is what you want worth the hardship to get it?

Even without pointing out that most people don’t even want it, I don’t believe “generative AI” is worth the destruction of our communities, our waterways, our world. Frankly, nothing is worth that, I hope you agree. And yet, that environmental cost is often ignored in the conversation with slopagandists. Which is one reason I call them ‘slop propagandists.’ Their talking points don’t engage meaningfully with any of the harms, yet they’re adamant about the so-called benefits. When a slopagandist does acknowledge these harms, they usually hand-wave them away as the cost of progress. But they do not engage with which part of society, of our communities, bears the brunt of that cost.

Which brings me to the thought that sparked this month’s musings: “genAI” is social injustice, digital colonialism. The entire industry is founded on inequity and theft, and runs on the backs of marginalized people who are being directly oppressed by financial burdens and environmental degradation. And, other than creating more wealth for tech owners, for what purpose? The most popular use-cases for “genAI,” for instance, involve deepfakes of children and women. This is the sort of thing people want to ruin the planet, and the lives of actual human beings for..?

This is unacceptable to me. If we truly are striving for a just world, where every human being has the chance to be happy and fulfilled, and to lead a dignified, meaningful life, then this should be unacceptable to all of us.

So, yeah, I know it’s hard. I’m asking anyway: please stop perpetuating harm in the world. Please stop using “genAI” wherever and whenever you can. Please be vocal about it. You won’t be alone. You aren’t alone. I’m in this fight with you. Many people are in this with us.

We are all in this together. It’s worth the hardship to do right by each other.

[Addendum: Okay, so you know how a broken analog clock is still correct twice a day? Well. The Catholic Church is still responsible for genocide, colonialism, religious strife and oppression; it allows too many people to justify misogyny and homophobia; it covers up sexual abuse of children. I abhor all of that. Unequivocally.
Yesterday, the Vatican released the latest encyclical letter from Pope Leo XIV: “on safeguarding the human person in the time of artificial intelligence.” On these points, I can agree on principle.]

Joyful Things
(because we need reminders that life includes joy, in ways big and small)

• I piggybacked onto my husband’s business trip to Vancouver earlier this month. I got to stay at a swanky hotel with a view of Coal Harbour. My best friends in the Lower Mainland introduced me to the best Thai food I have ever had in Canada. I met my mum for lunch on a different day and shared the best xiao long bao in the city. I read the latest Murderbot book without interruption, with the windows open to the sound of water and seabirds, and the occasional goose. I watched a beautiful heron fly by.

• I recorded my first interview for Season Two of the podcast last week! And yesterday, I recorded a second. There are two more in the scheduling hopper, and many invitations in the wild, awaiting replies. I’m so grateful people are feeling good about this show and are open to helping me put joy into the world.

• Last month, when I was searching for an IMAX theatre to watch Project Hail Mary, I discovered that The Thing (1982) will be playing at the end of May, on the IMAX screen at Telus World of Science (basically our planetarium+, for those unfamiliar). So, of course I immediately bought tickets. I was too young to catch its theatrical release (which critics panned! How dare!!), but I remember watching a VHS rental with my cousins and brother. We were all too young lol, but I loved it. I can’t wait to see it writ large in a theatre!
  (BTW I’ll be on a panel next month for the Incomparable Mothership podcast about Project Hail Mary. That show records well in advance, so it may take a while, but I'll share when I have a link to the episode.)

Oh, there’s just one more thing...

What’s bringing ✨you✨ joy these days? Please hit Reply and let me know! 🤓