Lee Schneider's Blog: 500 Words

Welcome to 500 Words, a letter about indie publishing, artificial intelligence, and the creative process.

Though I write about AI, I'm not an AI fanboy.  I'm interested in interrogating IA, which means I'm going to question all of its basic assumptions. The reason for this is that I am a recovering early adopter, meaning that once I was an enthusiastic grabber and tester of all new technology. But I've become more skeptical of the bright, shiny tech object, and I question the motives of the people who are pushing all the new tech on us. If that viewpoint works for you, consider subscribing. Every issue of 500 Words is free. You can also support my work with an optional paid monthly subscription, if you wish, or just buy me a coffee. 

INTERROGATING AI
This must have sounded good back in April 2020: A robot for kids named Moxie that was marketed as a safe and engaging animate companion, designed to promote social, emotional, and cognitive development.  Do you have a warm spot growing in your heart? Sorry, you'll have to do something about that in the next paragraph. 

Welcome to 500 Words.

When I worked as a news writer and producer, I acquired a habit that I am now trying to break. Back then, every day, I churned through as many print and internet sources as I could get my eyes on. Now, the more news I read, the worse I feel.

In the episode, I go on to talk about how I've tried to change my news input habits. 

In INTERROGATING AI, I cover Pokemon Go's data grab, Jerry Garcia's immortal voice, and the influx of VC funding into a startup that is developing autonomous weapons. 

Well, that didn't go well. A majority of Americans have voted for hate, choosing a criminal to lead us. It feels like a day, or a few days, or maybe a whole week to dip into despair or try to recharge. I have given some time over to despairing and mourning, and now I'm working on the recharge.

Can there be dialogue in our divided land? Is it worth it to try reaching across the aisle?

A moment for station identification for new readers and listeners:  500 Words is a weekly-ish newsletter with two regular columns: INTERROGATING AI, which examines our use of artificial intelligence, and BLURBOMATIC, a monthly curated list of book reviews. I'll also keep you updated about books I'm publishing, productions in progress, new courses and trainings, and podcasts I'm launching or publicizing.

I have made a point of writing about productive uses for AI, like using a chatbot as a writing coach to help you come up with ideas. Adobe has added a feature to their editing app, Premiere Pro, that will generate a few missing frames of video if yours ends too early to make the edit.  AI has the capability of making us fitter, happier, more productive (Yes, a that's a Radiohead quote from *Ok Computer*.)

But truly interrogating AI means asking who it benefits. Easy answer to that: The companies leading AI development have a beneficiary in mind--themselves. AI has corporate masters, and it is being shaped to benefit them. We, as users of AI, can speak up and make it more beneficial to humans, and we ask demand that it not hurt the Earth too much. 

There are many AI unknowns this week. To sort them out, let's follow the money.

OpenAI is restructuring from a non-profit to a for-profit company. Its Chief Technology Officer, Mia Murati, announced that she's leaving. Murati played a pivotal role in getting OpenAI CEO Sam Altman booted from the company last year, and she briefly became its CEO in his absence. Then, with Microsoft's financial backing, Altman regained control over the company. 

Whatever power struggles occurred behind the scenes, they have benefitted Altman. After the restructuring, OpenAI is discussing giving him a 7% equity stake. At OpenAI's current valuation, Altman's 7% would be worth around $10.5 billion. 

Hey, you human? Prove it. 

With more AI-generated slop that I can handle on my screens these days, I say it's time to issue online credentials that will prove that you're a human being. 

We have driver's licenses to prove who we are when we get behind the wheel, and we have passports to prove who we are when we cross a border. You need a photo ID to buy a gun. 

On the internet, you can be a dog, and nobody would know. I listened today to a couple of bots doing their own podcast, and they were pretty darn close to sounding human. Bots are writing articles, financial reports, and meeting minutes. Bots are also degrading the internet, spreading misinformation, committing fraud, and directing cyberattacks.  

AI's might seem pretty powerful, but there's one thing they can't do. An AI can't show up somewhere in person. So personhood credentials would involve going somewhere, like the DMV or the Post Office, and proving your identity, and somebody official there would issue an encrypted credential. The credential itself it could be a certificate in your web browser that would be linked to your fingerprint, eye, or voice.

Not only are people buying more books, but they are also talking about them more often, in more venues. BookTok helped sell about 20 million books in 2021, more than 2.4% of book sales for that year.  The #booktok hashtag has more than 200 billion views on TikTok, which is like saying every person in the world has viewed the hashtag 25 times. 

Confession: I've never looked at that hashtag in its home environment, on TikTok. I'm not interested in another blippy platform. I'm looking instead for longform conversations about books. Conversations like that are getting easy to find. 

Hello, and welcome to 500 Words.

It didn't seem that long ago that the experts said that nobody would buy physical books anymore. Well, Barnes & Noble is thriving, on track to open 58 new stores this year. That's the fastest pace the company has set for openings since 2009. Physical books are winning in the age of social media, aided *by* social media, like Facebook, TikTok and the popular #booktok hashtag. The new Barnes & Noble stores will be coming to New Orleans, Northbook, IL, Conroe, TX, Santa Fe, NM, Fayette, KY, and Santa Monica, CA, among other places. 

How can you be sure I'm me and you're you? I'm asking because a person's vocal identity is hard to verify online, even if you know their voice well. Synthetic speech generation has gotten so good, it's close to impossible to tag a voice as real just by listening to it. 

Earlier this year, New Hampshire voters received a robocall that sounded like Joe Biden. Biden's voice told them to “save your vote for November,” encouraging them not to vote in the presidential primary. Doesn't sound like something that Joe would say, and it wasn't. It was a deepfake of his voice. 

Let's consider actors' voices as uncommon instruments, ready to fit vastly different roles and build big emotions with the smallest nuances of expression. Voice actors are hired on productions because their voice can be heavy as gravel or light as a butterfly. These voices may be virtually transparent to you, as you hear them narrating commercials, reading audiobooks, playing characters in animation and video games, and playing roles in podcasts. Some are famous, like Claire Danes and Maggie Gyllenhaal voicing audiobooks, and some you'd never expect to be voice actors, like Seth Green (Scott Evil in the Austin Powers movies), who has built up a repertoire of more than a thousand voice-acted roles.

On August 12, the actors' union, SAG-AFTRA, closed a deal with Narrativ, an AI company, that allows union members to license their digital voice replicas for use in digital audio ads. 

Some actors are saying they have to fight the machine, and others believe they have to use the machine because the AI companies are going to clone them anyway. The read I got from friends in the acting community is that most working actors are against cloning their voices, because it will erode their standing as unique talents and ultimately reduce their income.