AI is Democratizing and Democracy is Famously Messy
I am not going to pretend to be an expert in Artificial Intelligence; you likely aren’t either. But that doesn’t mean we shouldn’t have a place in the discourse around generative AI and its integration into society. We have known this day was coming for at least three decades but most people were blindsided when AI and Large Learning Models hit the mainstream last year with the public release of OpenAI’s ChatGPT.
The democratizing force of technology is possibly the defining characteristic of the last twenty years of human history. It levels the playing field between ordinary people and institutions. With a bit of computing power, dudes in their basements can day trade stock options and foreign currencies like a hedge fund. With a smartphone and YouTube, random Nigerian tweens can make movies and trailers mimicking (maybe rivaling) those made by Hollywood studios. With the right inputs into a generative AI, people who can’t draw stick figures can now render gorgeous landscapes and intricate characters.
This is the democratizing force of technology but the problem with democracy is that bad actors freely wield equal power. There’s two particular ways AI is going to impact the near-future that I want to touch on today and likely return to: amplifying grifters and empowering scammers.
Democratization is a sword that cuts both ways. While technology allows the elevation of voices and groups that previously lacked access to the commons, it also enables ghouls to seize and build audiences that were unimaginable a generation ago. It would be impossible for a figure like Alex Jones (or any of his conspiracy-spreading copycats) to rise to their current level of fame and wealth without the internet.
Without social media Jones and his ilk would be the modern version of the John Birch Society cranks or UFO truthers pushing their pamphlets at the local gun show. Instead, they hold sway over an entire political coalition and may serve as kingmakers in the 2024 GOP Presidential nomination process. The era of social media brought us Jones. The coming era of AI is going to bring worse figures into the zeitgeist and allow them to produce, publish, and share content at an industrial scale.
We live in a Golden Age of scams that’s going to get worse. At this point, ninety-ish percent of the phone calls and about one-fifth of the text messages I get are overt attempts by scammers from various locales to defraud me. It’s likely the same for you. As we’ve discussed in prior newsletters, these scams are often carried out by large well-funded, organized syndicates. Running a call center in Myanmar with hundreds of people working is capital intensive. But AI tools are going to drastically lower the bar to entry into the scam game. Rather than needing large numbers of people in call centers or technical hacking expertise, off-the shelf tech tools are going to create an arms race between scammers and banks… and my money ain’t on the banks.
The video below shows how easy it is to clone someone’s voice while offering some very weak caveats about why doing it is bad.
Imagine a committed scammer using a language learning model to scrape the internet for all the available online information about you. Next, they use AI tools to spoof your voice and call your bank, using the information from the learning model to gain access to your account. When asked to show credentials, they can create realistic looking legal documents and IDs. If needed, they can create a full motion video of you for a Zoom call. Now, imagine the likelihood of the last bank teller or operator you talked to being able to suss any of this out.
The technology to do all of this is either already on the market or deep into development.
In the past, I have been imprecise about this: my concerns are not about the tech, which is (arguably) value neutral but about how we can protect ourselves from bad actors wielding the technology. This is a time for us to be figuring out societal safeguards. In particular, our elders and the less tech savvy among us are set up to be unfortunate marks. We are woefully unprepared for the ways people will leverage the current and next generation of AI and our policymakers are too busy trying to ban AP African American Studies and books about gender fluid kids to do anything about any of it.
See you next week, I’m going to change all my passwords, again.
As always, thanks for reading the newsletter. If you'd like to respond just hit reply on this email and if you like the newsletter, share it with somebody you love.
See you next week!
As always, if you have any thoughts or feedback about the newsletter, I welcome it, and I really appreciate it when folks share the newsletter with their friends.