My Condolences from ChatGPT
Why You Maybe Shouldn't Use AI to Reject Someone
Welcome to my second post for Peculiar People. I hope everyone had a great Easter. If you missed my introductory post, that’s a great place to start to get an understanding of what I’m trying to do in this space.
If you haven’t subscribed yet, you can do so for free. I’m not putting any posts behind a paywall for now. Also, if you subscribe, you can catch me on Substack Notes. (More on that at the end.)
Rejected by a Robot
I almost used ChatGPT to reject a book proposal. Don’t worry, I didn’t do it! (I can hear my authors and agents sprinting to check their inboxes.) But I was close.
As an acquisitions editor, rejections are my least favorite part of the job. Others may disagree, but it’s not very fun to crush someone else’s dreams. Yes, I’m being a whiney baby, but I find it all emotionally taxing.
Enter ChatGPT. When it burst onto the scene and took over the internet in what felt like only two minutes, I at first watched with mostly disengaged amusement. It seemed like a gimmicky piece of technology for those who wanted to see things like Abraham Lincoln’s Gettysburg Address in the style of Eminem.1
But one day I had a particularly difficult rejection to write and couldn’t quite find the right words for it. I suddenly became curious about how ChatGPT could write a gracious rejection letter. What it spit out was pretty dang good—better than what I could have written (which probably says something about me, but we can talk about that later). It was so good that I was tempted to just copy and paste it into the email.
Boom. Easiest rejection ever.
But then a thought stopped me. Is it fair to this person who is passionately pursuing their dream for me to reject them by a robot?
Me: But rejections aren’t fun for me.
Also me: They aren’t fun to receive either, you jerk. Wouldn’t it be better if they received it from an actual human being?
Me: Yeah, but they would never know.
[Pause]
Also me: But what would such an act do to you?
Me: [slams head on keyboard]
Discussions are happening right now about the dangers of generative AI (think Skynet and Ultron). And then, potential extinction scenarios aside, there are additional concerns about copyright infringement, disinformation, invasion of privacy, and the looming disruption to the labor market. Those are important discussions to have and I’m glad people far smarter than me are having them. But I think my main concern is a little less apocalyptic. I’m more worried about what adopting such technology might do to us.2
That rejection email was a moment where my humanness was needed to pass along disappointing news, but it was shocking how easy and thrilling it was to have that email written for me. Is generative AI just another piece of technology that will lull us to sleep, further separate us from other people, and continue us on our trajectory of making this world less hospitable to humans?
Superpowers
Andy Crouch in his book, The Life We’re Looking For, talks about how certain pieces of technology are basically like having superpowers. They promise “power without effort.” Take our smartphones for example. You can carry the entire knowledge of the world in a tiny rectangle that sports a Spongebob Squarepants PopSocket. Superpowers.
It’s all pretty awesome, right? Well...sort of.
Superpowers give us the ability to escape the built-in limitations of our humanity. And the constant use of those superpowers inevitably shapes us. The problem, Crouch writes, is that a superpower “diminishes us as much as it delights us.” We are escaping something that made us human and replacing it with something else. New technology brings with it new habits, and new habits bring with them new loves and new ways of being and thinking.
If you want to know how technology has shaped us, ask yourself how much the world has changed because of the smartphone. How much has your own life changed? How is your attention span doing? Your mental health?
When a new piece of tech comes our way, we should be asking, “How will adopting this new technology shape us as human beings? How will this shape society?”
Superpowered Communication
Microsoft and Google have announced that their respective chatbots will be built into their office apps, including email and word processors. That means you will get a true and better Clippy right there for you whenever you need to write another company-wide memo. But it doesn’t have to stop at emails. People have used chatbots to write cover letters, resumes, dating app DMs, sermons, wedding vows, and love letters. Some have even turned to their chatbot to guide them in how to talk to their boss, their kids, or their spouse.
The possibilities are endless. Don’t know what to say in a text message? Bing will help. Struggling to find the words for your niece’s birthday card? Bard is there. Have to write a new Substack? ChatGPT can do it for you in just a few seconds.
It’s helpful and convenient and feels like superpowers—imagine all the rejections I could get done in one afternoon!—but is it actually good for us? I’m not so sure.3
With the near-instant proliferation of chatbots, we are not far from a plausible future where we will have to assume that all written communication we receive will have been mediated by a chatbot. The majority of our communication is already mediated, either through text messages, DMs, and emails. What’s one more step of removal?
“That’s progress now, the steady accumulation of various tools to avoid other human beings.” Freddie deBoer
At first, that may sound kind of nice. “I don’t have to be the one to think up what to say anymore! It’s exhausting knowing people.” That’s true. People can be a lot of work. I’m an introvert and other people’s emotions scare me.4 Think of all the time and energy I gain with my new superpowers! But we should also ask ourselves: what do we lose?
My Condolences
As a test, I decided to ask Bing (Microsoft’s GPT-4 powered chatbot) to write a fake condolence text message to a fictional friend whose house burned down, along with all his possessions. Poor guy.
Just like the rejection letter, it churned out a pretty decent message, although a little clunky in parts. But it did the trick, imitating compassion and empathy. (I also had Bing’s AI image generator powered by DALL-E create the header image for this post.)
Bing then suggested that I could ask it to make the message even more personal. I, of course, clicked the button because I wanted my buddy to really know that I cared about him. What it spit out was even more heartfelt. I’m sure I could have sent that off to my fake friend and he would feel so loved by me.
But he wasn’t loved by me. I didn’t do anything. A gigantic program working through trillions of pieces of data did it in about 7 seconds. How would my friend feel if he found out who (or really, what) had written the message? Would he truly believe that “I’m always here for you, no matter what” when I couldn’t be bothered to write a text message on my own?5
From 2023 on, that note of condolence from your childhood friend or that termination email from your employer could have you wondering, did a human actually write this? This was recently demonstrated when Vanderbilt University was caught using an AI chatbot to write a letter to its students about a recent school shooting.
In the US there’s a loneliness and mental health epidemic, especially among the younger generations. Will this kind of mediated world help us or actually make us feel more isolated and separated? They’re even talking about having chatbots used as therapists now and creating personalized AI friends for kids.
It’s hard to do something like write an email to students about an awful shooting, but it’s precisely the sort of thing that we need to lean into with our humanity. Some communication is supposed to be hard. It needs a human touch, not necessarily because we might write it better than the chatbot (we probably can’t), but because people need to know that a human is on the other end of the line. But when we shortcut it, we show that the task—and perhaps, even that person—wasn’t worth the trouble.
Instead of turning to AI to write difficult messages for us, what if...hear me out on this...we are supposed to lean into the hard things? What if they are opportunities for us to be more human, training ourselves to better love God and our neighbor? We can’t run from the hard things, the human things, forever. To quote deBoer again, “That’s the existential question for humanity in the 21st century. That’s the challenge in front of all of us. Will you shoulder the risk of pursuing real human connection, as hard and intimidating and discouraging as that can be?”
Idol Factories
And yeah, I’m being rough on technology right now, but the reality is that the problem more has to do with who is creating and using the technology. The problem is us. Technology is not neutral because humans are not neutral. John Calvin wrote that our hearts are little idol factories. We are the problem. We are the ones making the technology after all, and far too often we are looking for something to make our idolatry easier. Like CS Lewis wrote in The Weight of Glory, “We are far too easily pleased.” We don’t often love what is best for us. We think we know what will satisfy us, what will make our life easier—escaping our humanity with superpowers, building a tower to heaven—but we usually pick the wrong thing.
Yes, I see the irony of my screed about technology being posted online by a computer, promoted by algorithms, and read on phones. I’m not saying we need to avoid technology altogether. I’m advocating for sober and careful evaluation, a healthy skepticism.
As Christians in a technological society, we don’t have to be complete Luddites who shut ourselves off from any technology. In many ways, it’s almost impossible to go against the grain and fully opt out of our techno world unless you form your own community separate from society. I have loads of respect for the Amish, but I’m not sure the majority of us need to get to that point yet. But being a peculiar people means we need to see beyond the allure of the new, the cult of productivity, and the altar of commerce. Let’s make sure that we’re evaluating everything, especially the things marketed to us by multi-billion-dollar corporations. As followers of Christ, it is incumbent upon us to be the ones to say, “Slow down! Will this help us be more human or less? Is this participating in the commodification and objectification of myself and my neighbors? Is this going to help us flourish?”
It took me an excruciating ten minutes (I know, poor me), but I was able to write out a genuine, heartfelt rejection letter. The hard, unproductive, and inefficient things are not a bug of humanity but a feature. We are not machines and we shouldn’t pretend to be.
A Small Thought
In about a month, I’ll be teaching an ethics class at my church to a small group of leaders. One of my favorite ethics thought experiments is known online as the “shopping cart theory.” The basic gist of it is: Do you put the shopping cart in the designated corral or not? And if you don’t put the cart back, does that say something about your character?
In my opinion, it’s a more fun thought experiment than the famous trolley problem,6 because it’s a low-stakes scenario that almost everyone faces...and it’s probably more complicated than you think.
What do we owe people? And why?
A Thing of Beauty
In honor of the recently passed Yang Bing-Yi (founder of the Taiwanese chain of restaurants, Din Tai Fung), I want to highly recommend that you eat some soup dumplings this week (xiaolongbao). If you’ve never had one before, it’s not dumplings in soup—although that’s a great way to eat dumplings—but a specific kind of dumpling with soup inside them, along with delicious meat and veggies. They’re an incredible invention, one of life’s true joys that proves that good food matters.
What I’m Watching
The Mandalorian
Duh.
I know some people feel this current season is not living up to their expectations, but I’ve been enjoying it. One of the best aspects of this season has been its greater focus on the devout followers of The Way of Mandalore.
Something about the Mandalorians reminds me of the Jewish people. Having been apocalyptically crushed by an imperial force, they’ve found a way to continue being a people amidst intense opposition and isolation. Little communities around the galaxy have sprung up, centered around forges that are the heart of their religion and practice. The creeds, the tales of old, and the prophecies of the restoration of their homeland, all combine to form a specific way of living and being in the galaxy.
Perhaps piety, liturgy, and spiritual disciplines are not dead and are the very things people need to see them through a chaotic world.
I’ll probably write on this more after the season is over.
Links
“You’re Better Off Not Knowing,” Shadi Hamid - The Atlantic
Maybe consuming so much news isn’t good for us. “Unless you have a job that requires you to know things, however, it’s unclear what the news—good or bad—actually does for you, beyond making you aware of things you have no real control over.” Major blow to us news junkies, but wise words.
“Most of the things we could know are a distraction from the most important things that we already know: family, faith, friendship, and community. If our time on Earth is finite—on average, we have only about 4,000 weeks—we should choose wisely what to do with it.”
“Why I Love Writing by Hand (And Hate Computers),” Jared Henderson - YouTube
I’ve been writing by hand a lot more lately. In fact, the first draft of this post was written by hand in a good old-fashioned composition notebook. I’m not sure if that made my writing better, but it was more fun to do it and it felt like tapping into something old and romantic about the act of writing.
“NASA Astronaut Asks for Prayer for Moon Mission,” Dan Silliman - Christianity Today
We’re finally going back to the moon next year, and the pilot of the mission, Victor Glover, is a Christian. In our modern world, we usually pit science and faith against each other. Here’s someone who’s able to reconcile the two.
Notes
I’ve been enjoying posting on Substack Notes, and would love for you to join me there!
Notes is a new space on Substack for us to share links, short posts, quotes, photos, and more. It’s like Twitter, but without the angry vitriol or He-Who-Musk-Not-Be-Named. I plan to use it for things that don’t fit in the newsletter, like work-in-progress or quick questions.
How to join
Head to substack.com/notes or find the “Notes” tab in the Substack app. As a subscriber to Peculiar People, you’ll automatically see my notes. Feel free to like, reply, or share them around!
You can also share notes of your own. I hope this becomes a space where every reader of Peculiar People can share thoughts, ideas, and interesting quotes from the things we're reading on Substack and beyond.
If you encounter any issues, you can always refer to the Notes FAQ for assistance. Looking forward to seeing you there!
Yo yo yo, check it out
We're here to honor the fallen ones
Who gave their lives for this nation
They fought for freedom and equality
And now we gotta keep up the dedication
Four score and seven years ago
Our fathers brought forth a new land
Conceived in liberty and democracy
And now we gotta make a stand
I hope you see all my “mights,” “coulds,” and “maybes.” I want it to be clear that these are concerns and not necessarily predictions.
Now, don’t get me wrong. There are lots of potential benefits to generative AI. Since GPT-4 can read text and also see images, OpenAI has partnered with an organization to help visually impaired people use GPT-4 to help them. There are interesting implications for expanding the ability to learn new languages that better resemble natural speech (instead of the outdated staleness of typical language learning apps and textbooks). It can be a tutor which to help low-income and impoverished communities. I’ve heard it’s not a bad meal planner. And there’s a whole host of other possibilities. Generative AI has the potential to do a lot of good. But at the same time, we must be careful and ask ourselves what sort of guard rails are needed for the programs themselves, and for us too.
I’m working on it.
The chatbot also didn’t have the wherewithal to suggest that maybe a text message isn’t the best medium for this kind of conversation. Perhaps a phone call at the very least?
Although, the show The Good Place found some ways to make the trolley problem fun…or at least entertaining.