Latest / HumanPrint: How to Use AI Without Losing Yourself / Episode 11: How Do You Sound Like You?

Transcript

Episode 11: How Do You Sound Like You? · HumanPrint: How to Use AI Without Losing Yourself

• 0:00
  Foreign.
• 0:05
  Welcome to Human Print. I am Christine Whitmarsh, and if
• 0:09
  you are open to collaborating with AI but you're not interested in disappearing
• 0:13
  into AI, you're in the right place. Each week, we will look at
• 0:17
  one place where your authentic human signal is at risk for drift
• 0:21
  and one way to shore it up. Oh, and Human
• 0:25
  Print is brought to you by Abby AI Brand bank
• 0:28
  intelligence that helps authors, thought leaders, and enterprise teams
• 0:32
  train AI on their voice and point of view so written content
• 0:37
  stays unmistakably human, consistent and on brand.
• 0:41
  Learn more at abbey.incabbi.ink
• 0:47
  like you write with. Let's begin.
• 0:52
  Welcome back to Human Print. Today's episode
• 0:56
  is called how do you sound like youe? You know, every time
• 1:00
  I see one of those social media posts, you know, here's how to know if
• 1:03
  AI wrote it. My response, usually internally, is the
• 1:07
  same, how do I know if you wrote it? Before you
• 1:11
  start accusing everyone else of using AI, can you actually identify
• 1:14
  what makes your own writing recognizably yours?
• 1:19
  So, yeah, welcome to Know Thyself Writing Edition.
• 1:23
  People tend to approach this conversation backward. They collect
• 1:26
  a list of phrases they associate with AI and they start cutting them out.
• 1:30
  Certain transitions, punctuation, sentence patterns,
• 1:33
  whatever. The AI Gotcha police have decided this week is proof of bot
• 1:37
  activity. And even if they do catch every supposed tell,
• 1:41
  which they probably won't, they're left holding a more important question.
• 1:45
  What remains? Removing generic language doesn't
• 1:49
  automatically reveal your voice. You have to know what you're
• 1:53
  trying to preserve in the first place. Spotting AI writing, it's getting
• 1:57
  easier and easier, especially when someone's pasting in raw, unedited output. We know the
• 2:01
  tells. These systems are trained on enormous amounts of human language,
• 2:05
  and they generate text by modeling statistical relationships within
• 2:09
  that data. When they have no real information about you,
• 2:13
  well, they're just going to generate random things from that big generic mixture,
• 2:16
  that random mixture of words that's not automatically your identity
• 2:20
  just because you attached your name to it. Back in the early chatbot days
• 2:24
  a few years ago, when I asked a generic LLM to write something for me
• 2:28
  as an experiment, I was the part that disappeared me the
• 2:32
  writing. It might have been clean and grammatically correct and reasonably well structured,
• 2:35
  but it just didn't sound like me because the system had almost nothing to work
• 2:39
  from. Before I built my own voice intelligence system, custom bots trained
• 2:43
  around a person's actual writing voice and point of view. The tool was just pulling
• 2:47
  from generic, you know, general patterns. It needed to be working from something,
• 2:51
  and it was working from nothing. It needed to be working from me, not from
• 2:55
  just a big data set, which meant I had to understand what
• 3:00
  I sounded like on the page. Fortunately, I've been
• 3:03
  writing forever. And I've been ghost writing books since 2001,
• 3:06
  which means I've spent decades studying how other people think, speak,
• 3:10
  speak, tell stories, explain ideas, argue,
• 3:13
  joke, relate to a reader, and move through sentences.
• 3:17
  At its best, ghost writing is not collecting someone's favorite
• 3:21
  words and sprinkling them into a manuscript. A voice operates
• 3:25
  on a bunch of levels at once. There's words and grammar, but there's also rhythm,
• 3:29
  sentence length, layering, directness, emotional range,
• 3:32
  how people use examples, humor, questions,
• 3:36
  metaphors, stories, certainty, contradiction, pacing, how the writer
• 3:40
  positions themselves. Aren't you glad you're not a ghostwriter? Voice reflects
• 3:44
  identity, too. When I zoom out in my own writing,
• 3:48
  I see the curious scientist, the lifelong creative, the storyteller,
• 3:52
  the writing coach, the habits enthusiast who keeps wandering into psychology,
• 3:56
  someone who resists black and white conversations and loves language just
• 4:01
  enough to play with it all the time. That is way too much to cram
• 4:04
  into a catchphrase. Everything I just listed and more shows up in the cumulative
• 4:09
  experience of reading my work. I like to think so, anyway.
• 4:13
  When I read something I wrote years ago, I see myself.
• 4:16
  I see the show Don't Tell scene setups that come straight out of my screenwriting
• 4:21
  background. I see the parenthetical side thoughts because apparently one thought
• 4:24
  at a time would be too orderly. I see the dialogue and the air quote
• 4:28
  thoughts, the alliteration groups of three. AI did not invent those,
• 4:32
  by the way. They've been around forever. Groups of three, the dorky humor,
• 4:35
  metaphors everywhere. Mostly I write just like I'm talking
• 4:39
  to you right now. The form changes. You know, academic research needs a
• 4:43
  different register than a podcast in fiction, different than a business article.
• 4:46
  But the person, the identity underneath the writing doesn't change.
• 4:50
  So how do you start finding that person in your own writing without turning it
• 4:54
  into some big technical voice mapping project? Well,
• 4:57
  let's just start with three things for the purposes today, anyway.
• 5:00
  Rhythm, story, and humor. Let's start with rhythm at
• 5:05
  the simplest level, and I'm going to use some musical terms here. Do you write
• 5:08
  staccato or legato? I love describing writing this way.
• 5:12
  Staccato leans on short, separate sentences, each one landing
• 5:16
  with its own stop. Used well, it feels very forceful
• 5:20
  and blunt and fast and, you know, very readable, easy to read.
• 5:23
  Legato connects through longer, layered sentences that let the writer
• 5:27
  qualify, elaborate, interrupt themselves, build momentum
• 5:31
  before arriving at the point. Most. Yeah, mo. You know, most writers,
• 5:35
  they use a combination of both things. The real question is, which one feels most
• 5:39
  natural to you, staccato or legato? And what happens to
• 5:43
  it when you get self conscious or nervous or under pressure or overly
• 5:47
  formal, or you lean on generative AI? What happens to that favorite
• 5:51
  form of rhythm that you've established? Now, I'm fundamentally a legato
• 5:55
  writer. My thoughts build through clauses and side comments and qualifications because that's actually
• 6:00
  how I think and talk. And then I'll drop in a short sentence to land
• 6:03
  the point. Sometimes, usually the second my
• 6:07
  writing turns into a stack of short, equally weighted sentences, I know it's not
• 6:11
  me anymore. And believe me, this is just scratching the technical surface. There is so
• 6:14
  much more here. I just grabbed a starting place for our conversation today.
• 6:18
  So pull up a few pieces that you wrote before generative
• 6:23
  AI existed, ideally the ones that you still like. And you read them out loud
• 6:26
  and they sound like you. Where do you breathe? How long do
• 6:31
  you stay with a thought? Do your sentences build, snap, wander,
• 6:35
  circle back, move straight through? How do they move? Do you drop
• 6:39
  a short sentence after a long one to reset the rhythm? Now, this isn't a
• 6:42
  contest between long and short. You're looking for your own natural rhythm.
• 6:46
  Second, and here's a story and a personal example.
• 6:50
  I've told authors for years that readers need a reason to care
• 6:54
  who you are before they care what you think. That doesn't
• 6:58
  mean every book turns into a memoir. It means even a knowledge based
• 7:02
  book usually benefits from lived experience or a scene or a client's story.
• 7:05
  A moment that shows the reader why this subject is yours. I think
• 7:10
  there's an element of memoir in almost every book. Every, every good book.
• 7:13
  Anyway, without that useful information goes emotionally flat.
• 7:17
  The reader gets the lesson and they feel nothing about the person delivering it.
• 7:20
  Like when someone posts raw AI output on social media because it
• 7:24
  sounds polished or professional. The information might be accurate, but there's really
• 7:28
  no human in it. So I become less interested. I don't know who you are
• 7:32
  or why you care or why I should hear this idea through instead of
• 7:36
  through you instead of all the thousands of other people who have the same idea.
• 7:40
  So look at how you use story. That's the second thing
• 7:44
  of the three that we're talking about today. Look at how you use story.
• 7:47
  Do you open with a scene and pull the idea out of it or make
• 7:51
  the point first and back it up after? Are your Stories detailed and atmospheric
• 7:55
  or brief and functional, do you put yourself inside them or keep a little
• 7:59
  distance? What kind of details do you notice? All of that,
• 8:03
  that story and everything we just talked about. That is voice. And third,
• 8:07
  humor. Now use humor, which I think some people demand
• 8:11
  of their chatbots use. Humor is not useful advice on its own because humor
• 8:16
  is intensely personal. And it's a pretty broad range, too.
• 8:19
  Is it dry observation? Exaggeration? Wordplay?
• 8:23
  Self deprecation? Absurd specificity? Sarcasm? Dialogue? And when
• 8:28
  you do reach for humor after a hard point, to build rapport, maybe with the
• 8:31
  reader or to, you know, kind of lighten the tension, just to challenge an idea
• 8:35
  without preaching defensively, how do you do it? Like when you reach for humor,
• 8:39
  how do you use it as a tool? The goal is noticing
• 8:43
  how humor shows up when you are actually being yourself.
• 8:46
  Voice also involves point of view, which is bigger than having opinions.
• 8:51
  How do you handle uncertainty? How fast do you land on a conclusion?
• 8:55
  Do you like clean answers? Or do you naturally complicate a thing right after you
• 8:59
  you've stated it? How hard do you push the reader? Do you teach through instruction?
• 9:03
  Questions? Story comparison? Evidence?
• 9:06
  What claims kind of make you flinch? Or as one client
• 9:10
  of mine says, itchy? What does your writing say about what
• 9:14
  you actually value? All of these questions and more.
• 9:18
  These are the things that move you past the surface. Stuff that people obsess over
• 9:21
  when they're hunting for gen AI on social media.
• 9:24
  And what do they actually get from that hunt, by the way?
• 9:27
  The gotcha police? Maybe a sense of control during a period of huge
• 9:32
  technological change, you know, resistance to new technology. It has
• 9:35
  a long history, and it's rarely just ignorance or panic. Like some people are saying,
• 9:39
  new tools, they threaten livelihoods, status, identity,
• 9:43
  the entire distribution of power. So, yeah, people respond to
• 9:47
  real risk and imagined. Generative AI deserves criticism,
• 9:51
  research and governance. But accusing strangers of using it does
• 9:55
  nothing for your own writing, for the discovery of your own writing voice.
• 9:59
  Knowing your voice does.
• 10:02
  AI can do structural supplementation here. It can compare samples,
• 10:06
  spot patterns, ask you questions about your own choices, or test whether
• 10:10
  a new draft still sounds like the old you. But it cannot
• 10:14
  construct the first version of your identity as a writer. Now,
• 10:18
  here's a line that I like to use. Decorating a house that you designed
• 10:22
  versus building a house that you had no hand in.
• 10:26
  In the first example, the foundation and the layout and the major calls,
• 10:30
  they all came from you. And Genai helped you examine and
• 10:34
  extend what's already there. But in the second, you show up after the decisions are
• 10:37
  made, and you try to convince yourself that it feels like home. And I think
• 10:41
  I flipped those, by the way. You get the point, though, right? If you build
• 10:45
  the house, you're making the calls from day one.
• 10:48
  If you're just decorating it, you're not. How about that? So here's
• 10:52
  some work. After this episode, pick three pieces of writing from
• 10:56
  before you start started heavily using AI for those of you that do,
• 11:00
  pick the ones that still sound like you read them out loud. Make three columns.
• 11:04
  Rhythm, story, humor. Under rhythm, what your sentences actually do. There's that
• 11:08
  staccato legato bit. Under story, just jot
• 11:12
  down some notes. How do you use life experiences? And how do you write scenes?
• 11:15
  And how do you insert details into your writing? Writing. And under humor,
• 11:19
  what makes it sound human and specifically yours, what kind of humor do you like
• 11:23
  to use? Then pick a recent piece, maybe one that you had
• 11:27
  some AI help with, and compare it and just see how they match up.
• 11:30
  Don't start by hunting for proof that AI wrote it. Hunt for proof
• 11:34
  that you wrote it. All right,
• 11:38
  that's it for today's episode of Human Print. Until next time,
• 11:41
  use the tool. If it helps, just don't hand it the pen.
• 11:54
  And that's it for this episode of Human Print. If you want
• 11:57
  a practical framework for protecting your voice amid the AI noise,
• 12:01
  please follow the show. And before you go, take a minute and ask yourself,
• 12:05
  as the AI noise grows louder and louder, which areas
• 12:09
  of yourself do you need to protect the most?
• 12:12
  I'll see you next time.