GeoLegal Weekly #65 - Is AI writing more compelling than human writing?

I wrote a first-person book and launched an AI-generated newsletter simultaneously. Here's what I learned.

The LA Times Festival of Books attracts nearly 200,000 people who brave one of the most inconvenient traffic geographies known to humankind to celebrate the written word. In a time of AI generated content and microscopic attention spans, mobilizing hundreds of thousands of people to celebrate books is a stunning achievement.

I spoke on an AI panel alongside Steven Witt (whose new biography of Jensen Huang sounds amazing), Brian Merchant (whose seminal book about the Luddites is more relevant now than ever before), and we were moderated by Joanne McNeil (author of Lurking: How a Person Becomes a User). You can watch the full panel here starting at 5h15mins.

It should be no surprise that when the topic turned to AI-generated content, my journalistic co-panelists seemed pretty skeptical that there would be a market for it. One of them noted that no one wants to buy tickets to watch two computers play chess against each other and another asserted that AI can only mimic and remix not produce genuinely creative content.

This helped me frame something I’ve been living for the last few weeks. My book, UNRULY, came out in late March and a few days later I launched a completely AI generated daily update on global affairs (produced by our new Hence Global platform). Perhaps I’m trying to be old school and new school simultaneously, or intellectual and anti-intellectual all at once. But the experience of launching a book and an AI-powered newsletter in synchronicity gives me something to compare and contrast.

1000+ People Signed-up to Read my AI-Generated Newsletter in One Week

What’s interesting is that I’ve been writing this Substack in a personal tone with my deepest thoughts for the last 64 weeks and am still a bit shy of 2,000 followers. My new Unruly Daily newsletter, which is almost 100% AI generated, scaled to 1,000 subscribers in one week and continues to grow. That is 1,000+ people who want to see AI generated content show up in their mailbox secured in days vs 2,000 followers over 18 months for my longer form writing. I’ll try not to take it personally!

Yes, they are on different platforms and that may account for some of the difference. But let’s dig a little deeper on what’s going on here and what we can learn about it as professionals. Our new Hence Global AI platform is designed to produce a personal daily briefing that is tailored to your interests and exposure. If you are the employment lawyer of a law firm, you’ll see a different daily update than an intellectual property lawyer. Because even though you live on the same volatile planet, the things that matter to you and your colleague are different.

So I told Hence Global that I wanted a personal daily briefing that categorized news based on the three synthetic risks I talk about in my book UNRULY (Geolegal Risk, Artificial Politics Risk, and LegalAI Risk).

There’s lot of risk in the world so the system does a good job cataloguing and categorizing it. I simply take that output and publish it as a LinkedIn newsletter for 1000+ people. You can take a look below at an example.

Before thinking about what value the written word has in a sea of AI generated content, I want to tease out a few conclusions about what the AI is good at:

1. AI content can be valuable and engaging if it’s actually relevant and intelligent. Nobody wants to read garbage from humans or computers. So, if the AI understands YOU and everything YOU care about, then it’s going to do a good job of serving you things you want to stop to read. And other people may share your interests so it can be the basis of publications.

2. AI content is much easier to produce - I would never have signed up to write a daily without it. So, perhaps, subscribers want some of my brain every day and are happy to have it be AI modelled on my framework.

3. AI content can reflect a broader corpus of knowledge than experts ever could. Our system can digest all of the world’s news in English and foreign language nearly instantly, while I can barely speak Spanish after ten years of studying it and would have to devote my entire morning to reading the sources that matter for what I have to do. The point is, on many things the AI is smarter than me because it can draw conclusions from a broader pool of data than I can.

4. AI may not be able to be the next Van Gough or Vin Diesel but it could be the next DJ Khaled. Which is to say that AI may struggle to create great, novel art but it could be better than mediocre at remixing existing art. If you had complete visibility of all art that came before you, your ability to generate something passably arty by combining it would be pretty high. It’s the same with political science - AI may not come up with the best new political science theory but if it has full awareness of all the theories put forward in history, it may be able to pattern spot the theory that matters the most to the moment we’re living in, generating a seemingly novel insight.

What role for humans?

OK so I’ve probably already alienated myself from the rest of the Author’s Guild by becoming a producer of AI-generated content, but let me try to work my way back into their happy hours.

Since I released my book UNRULY I’ve been really surprised at how much the personal stories within it resonated. Most books written by analysts are written in third person with an all knowing voice asserting truth after truth. On some level, that’s how we’re taught to write - it’s meant to be more formal and more convincing. Much of UNRULY describes truths about the world and provides general strategies for winning within it.

And, yet, most of the feedback I get about UNRULY is related to the personal stories I share within it. I talk about fighting an insurance company with AI, having my car stolen, showering with a bathing suit on in China, and, getting the f-word yelled at me by hedge fund clients (a popular passage with my young children). While I may think that that the legacy of the book is the big concept on which its hung, actually the legacy of the book is an encapsulation of my personal experiences and voice in a way that makes readers either want to get to know me more or head for the hills.

And, that reaffirms to me the core strategy for fighting back against an AI generated world. First person storytelling is perhaps the hardest thing for AI to emulate because AI is not a person and does not live human experiences! To be sure, AI can spin yarns with the best of them. But my goal here is to compare AI truths with human truths; anything else would be like comparing books that are big fat lies with books that are not.

As third person narratives become indistinguishable with what AI can produce, I suspect we’re going to see more first person analysis that connects personal experiences to what is going on in the world. Situating the author’s authority and lived experience in whatever topic they are projecting expertise about will help distinguish it from a world of AI generated content.

So as a business leader, how you can take advantage of this:

• First, don’t be timid in a world where having a voice seems like a liability. Figure out what your values are and figure out how to articulate them with authentic stories from your past. This will help you break out of the AI-generated cesspool that is all other AI-generated content except for my newsletter (just joking…).

• Second, think about how AI can make you a more efficient writer. I don’t think there’s any shame in that. AI can help you identify historical parallels that are relevant, choose case studies, test argument strength and monitor the world for up to the date examples.

• Third, remember to personalize and contextualize. There’s nothing I hate more than AI spam in my mailbox. But it’s not the AI that’s the problem. Humans have been spamming for a long time. AI tools that understand what the target audience wants to read can be used effectively - perhaps even more so when combined with authoritative personal voices woven in.

A bit different from my normal post but hope this makes you think - please let me know if you agree or disagree in the comments! You can even have AI generate them for you.

-SW