WR 1: Welcome to the Weekly Reckoning!
WR 1: Welcome to the Weekly Reckoning!
A new weekly edition of the Ethical Reckoner. Today: the week of 2/10/23
Welcome back to the Ethical Reckoner. You might have noticed that things look a little different around here, because today, I’m debuting a snazzy new logo and a whole new format!
This is the first installment of the Weekly Reckoning, the Ethical Reckoner’s weekly edition. Don’t worry, there will still be a full-length ER every month (or so…), but I wanted to introduce a weekly edition to bring you more timely topics and discussions. Every week, there will be a three-parter* in your inbox.
The Reckonnaisance: Stay in the loop on the week’s most important tech news. I’ll curate a few of the week’s top headlines and explain why you should care about them. (For a more in-depth news debrief, check out the Tech Ten from Josh Cowls at plus/minus.)
Extra Reckoning: Every week, I’ll spotlight a topic that’s caught my attention. I’ll offer a few links and dive into what I’ve been thinking about them—sort of a mini-ER, but more exploratory.
I Reckon…: a closing thought or question that I’ll be mulling over for the next week, and invite you to contemplate and discuss.
*Format and names subject to change. Terms & conditions may apply. Etc…
The Reckonnaisance
Generative AI advances anthropomorphize LLMs even more
Nutshell: ChatGPT now can talk to you in shockingly realistic voices, and Meta has launched AI “characters” that you can message on Messenger, WhatsApp, and Instagram. (If you speak English.)
More: Some of these characters are fictional (like “warm-hearted grandpa” Brian), some are based on real but dead people (like Jane Austen), and others are based on real people (like a “Dungeon Master” based on Snoop Dogg, Billie the “Ride-or-die older sister” based on Kendall Jenner, and “travel expert” Lorena based on Padma Lakshmi).
Why you should care: In a previous ER, I objected to giving LLMs real names because it unhealthily personifies algorithms. This ramps it up to 11. The “character” conceit seems to be intended to provide some distance from the real celebrity the “familiar face” characters are based on, but I’m not sure how much separation there will actually be. Presumably, these have been trained on their writing and/or transcripts of their speech, and the characters’ Instagram accounts feature real pictures of the celebrity (plus, there seems to be a generated “live” video of them when you’re chatting). This seems like a recipe for sketchy parasocial relationships—people will start feeling like they know the celebrity personally, which is what leads hundreds of people to swarm a rehearsal dinner that Taylor Swift is at. ChatGPT doesn’t have a defined personality, but the voice is so good that, as Hard Fork said, “people are absolutely going to fall in love with ChatGPT.” People are already in love with Replika chatbots, and giving AI a name, personality (especially that of a celebrity), and/or a voice means that relationships with AI are going to become even easier, with potential real-world impacts on the people interacting with these chatbots and the celebrities lending them their faces and personas.
Being female online is getting even harder
More: The fake images were created using clothed pictures taken from the victims’ social media accounts. Over 28 girls have been impacted, with severe impacts on their mental health. Eleven boys are being investigated, and some even tried to extort the girls for money. Spanish law isn’t equipped to deal with the generation of sexual images for adults, and it’s unclear if it falls under child sexual exploitation material laws.
Why you should care: Now, any woman or girl with a few photos of herself online can be personally humiliated with very little effort. Laws may not be equipped to deal with this and the fact that in many cases it’s being used to generate images of underage girls is horrific. It also shows why we can’t rely on the good graces of bad actors to prevent dystopic technological outcomes. China is mandating that generated images be watermarked, and some Big Tech companies have agreed to watermark their generated content, but the questionable efficacy of these watermarks and the proliferation of open-source models that anyone can install on their computers may mean this is our new normal.
The Supreme Court will rule if states can prevent social media platforms from moderating content
Nutshell: Florida and Texas enacted laws preventing social media platforms from “censoring users based on their viewpoints.” Now, the Supreme Court will rule on their constitutionality.
More: Platforms hold that these laws violate their First Amendment rights, since the laws are essentially the state forcing them to endorse and platform speech that they may not agree with. The states say that conservative voices are being unfairly censored. Appeals courts have conflicting rulings: one upheld the Texas law, while another struck down the Florida law.
Why you should care: A ruling for the states could prevent social media companies from moderating content more broadly, including removing hate speech and extremist content. Social media platforms would become pretty bad places to be (think Twitter today but even worse), and this would disproportionally impact people who are traditionally the targets of harassment and abuse.
SBF goes on trial
Nutshell: Disgraced crypto wunderkind Sam Bankman-Fried’s trial began Tuesday with jury selection.
More: SBF managed to incinerate $8 billion in customer funds when his crypto exchange, FTX,1 and its sister hedge fund, Alameda Research, imploded last November. He claims it was an accounting error; prosecutors say it was fraud. There’s legal intrigue—a secret backdoor into FTX’s systems? Illegal campaign contributions?—and it’s a fascinating look into a wild industry (where else can you work that will buy you a $30 million resort penthouse in the Bahamas?) but it’s also a juicy people drama. SBF was crypto’s golden boy until he wasn’t, and several FTX/Alameda executives, including his ex-girlfriend Caroline Ellison, have pled guilty and are cooperating with prosecutors.
Why you should care: Possible comeuppance for someone who exemplified the “tech boygenius who can do no wrong” trope. Crypto enthusiasts are hoping it will be a sage-burning for the industry, a “catharsis” for the crypto ecosystem. For everyone else, it’ll be popcorn with a shake of schadenfreude.2
Extra Reckoning
I’ve been thinking about what the Internet was before the Internet. I’m betraying my age, but having been born into a world where the Internet always existed, I’m fascinated by BBSs and other ways people interacted in cyberspace before they had the Internet. I really enjoyed The Modem World by Kevin Driscoll, which talks about the rise of Bulletin Board Systems (BBSs) in the 1970s and 1980s. Before the Internet as we knew it existed, cyberspace was a series of small-scale, text-based BBSs accessed via dial-up modems via small PCs or even “teleprinters,” which basically printed text instead of displaying it on a screen. People posted and had discussions with other people, similar to forums today. They started as spaces only for techies, but more and more people started using them—although only a small fraction of those who would eventually access the World Wide Web.
What intrigues me about BBSs is their particular form of community. They were run by sysops who were essentially (mostly) benevolent dictators. Sysops were “gatekeepers and moderators,” setting the tone for their BBS, but they also sometimes organized in-person meet-ups, which goes against the narrative that early techies pushed that everyone was disembodied online and that this was a good thing that created total equality. Sure, in a text-based BBS, you could generally conceal your physical identity (although some characteristics could seep in through the way you wrote), but people pretty consistently tried to figure out aspects of your identity (which we see through the old “A/S/L?”3 greeting in chatrooms) and enjoyed in-person meet-ups. And of course, there was also racist and sexist harassment when people discovered aspects of others’ identities, showing that you could only be truly disembodied as a white, straight, cisgender man, which most of the techies were.
There’s definitely a disconnect between the narrative of disembodiment and the desire to be more embodied online. People formed affinity BBSs centering around race, gender, and sexuality, and as soon as graphics came around, people started using them to create avatars for themselves. There’s a consistent trend towards more and more embodiment online—and sometimes tensions between what we say we want (or what the techies developing things say we should want) and what we actually want.
I’m a big believer in the idea that it’s important to look to our past in times of change, and we’re at one in terms of social media—with the drawn-out demise of Twitter4 and the proliferation of replacements—and potentially in terms of computing paradigms. There’s a trend towards becoming more embodied online, and VR could potentially make us fully embodied in cyberspace. Perhaps what’s coming is more small-scale social media, local and centered around shared interests. Or maybe it’ll be a mix of small and large (ideally outside the remit of the modern-day Big Tech platforms, as Ben Tarnoff’s Internet for the People argues). But I have to wonder what the role of the most-embodied tech platform yet will be. Yes, the “metaverse” and VR headsets are not great… right now. PCs probably weren’t great when they were first released. It took a long time for everyday people to adopt them, and we’re clearly in an early-adopter stage for VR right now. Apple’s new headset, although geared towards augmented reality (AR), may move the needle, if it’s good and they can make a cheaper one. On the other hand, maybe there’s a threshold for how much embodiment we want online. Platforms certainly still haven’t figured out how to deal with harassment that becomes even worse when it’s directed at your virtual person rather than through a screen. But maybe the future of social media is like the past: hanging out with your friends and other people who share your interests, in person because otherwise you had no way of contacting them, except “in person” will be in a virtual world.
I Reckon…
…that those of us who grew up with a semi-embodied, graphical Internet relate differently to it than people who grew up with a mostly text-based Internet, and that anyone who grows up with a fully-embodied Internet will relate to it even differently. If VR becomes widespread, how will different generations use it differently?
A website where people could buy and sell magic beans cryptocurrencies, like Bitcoin. For a good explanation of what happened, check out this piece from Intelligencer.
Although let’s not forget that there were 9 million real people who got snookered and lost much of their savings.
“age, sex, location”