Latest / HumanPrint: How to Use AI Without Losing Yourself / Episode 8: Words as Pixels of the Human Experience

Transcript

Episode 8: Words as Pixels of the Human Experience · HumanPrint: How to Use AI Without Losing Yourself

• 0:00
  Foreign.
• 0:05
  Welcome to Human Print. I am Christine Whitmarsh,
• 0:08
  and if you are open to collaborating with AI but you're
• 0:12
  not interested in disappearing into AI, you're in the right place.
• 0:16
  Each week, we will look at one place where your authentic human signal is
• 0:20
  at risk for drift and one way to shore it up.
• 0:23
  Oh, and Human Print is brought to you by Abby AI
• 0:27
  Brand bank intelligence that helps authors, thought leaders,
• 0:31
  and enterprise teams train AI on their voice and point
• 0:35
  of view so written content stays unmistakably human,
• 0:38
  consistent and on brand. Learn more@abbey.ink
• 0:42
  a b b I.ink
• 0:47
  like you write with. Let's begin.
• 0:52
  Hi, everybody, and welcome back to Human Print.
• 0:56
  Today, I want to start with something I wrote in an article about data
• 1:00
  storytelling. It's one of my passions because it's the idea
• 1:03
  that this whole episode today is built around. And that idea is
• 1:08
  each data point represents one
• 1:12
  pixel of the human experience. Now, in statistics, a data point
• 1:16
  is a number, but that number that comes from somewhere. It comes from a person
• 1:20
  answering a question or making a choice, experiencing a symptom or
• 1:24
  a behavior, buying something, leaving a job, showing up,
• 1:28
  not showing up. That number is the record. But the human experience is
• 1:32
  what generated it. And if you're only looking at the trend line, only looking at
• 1:36
  the pattern in aggregate, you can miss what's actually happening at the individual
• 1:40
  human level. And I think words work exactly the same way.
• 1:44
  The word is the pixel. The human experience that generated
• 1:48
  it is what gives it meaning. So AI can produce words, a lot
• 1:53
  of them, fluently, convincingly, but without the human experience underneath them,
• 1:57
  which means the pixels are there, but the picture might
• 2:01
  not be. Now, you know, I have my master of science in quantitative
• 2:05
  psychology. That's basically stats and human behavior. It's a combination of statistics and human behavior.
• 2:10
  I just love that intersection, by the way. It's one of those things,
• 2:13
  you know. So one of the things I came out of that program with was
• 2:16
  the ability to look at a scatter plot and kind of hold two ideas at
• 2:19
  once, the trend line, line and the individual dots. The trend
• 2:23
  is exciting. You know, patterns and data are exciting, but so is the actual
• 2:27
  scatter of the dots, because every dot in that scatter plot is a person with
• 2:31
  an experience. And sometimes the most important thing in the data set isn't the
• 2:35
  trend at all. It's the outlier, the dot that doesn't follow the pattern, the survey
• 2:40
  response that makes you go, wait, hold on. Why did this person answer this way?
• 2:43
  What did. What are they telling us? That the trend can't.
• 2:46
  So when I was running my scoliosis pain and exercise pilot
• 2:50
  study, it was my master's thesis. It ended up being called the SE18 Scale
• 2:54
  Project. It's on my website,
• 2:56
  christinewhitmarsh.com you know, the trends were genuinely exciting to
• 3:00
  me. The exploratory factor analysis, the pain variables that held it all together.
• 3:04
  But equally exciting were the individual survey responses.
• 3:08
  Because reading through them, I was seeing my own fellow scoliosis
• 3:13
  survivors, people who share a condition that I live with in my own body.
• 3:16
  And every. Every single response was a pixel of a human experience
• 3:20
  that I recognized. The trend told me one story. The individual
• 3:24
  dots told me dozens of others. And that's the macro and the micro at
• 3:29
  the same time. And I think that's what human judgment does, that pattern recognition alone,
• 3:33
  like AI cannot. AI is very good at
• 3:37
  the trend line, the pattern recognition. It's been trained on enormous amounts of
• 3:40
  data, and it's genuinely impressive at revealing patterns. But it doesn't
• 3:45
  understand the scatter. It doesn't understand the individual dots.
• 3:48
  It has the averaged out shape of a lot of human experience,
• 3:52
  which is a different thing than the specific experience of any one
• 3:56
  human when you're writing as yourself. And that difference is the whole game.
• 4:01
  All right, so bringing this back to book coaching, because this is where the pixel
• 4:05
  idea gets really concrete.
• 4:09
  Sometimes a client sends me a chapter and says something's off.
• 4:13
  It's technically accurate, well organized, the argument makes sense, but something
• 4:18
  is wrong and they can't say what. And the thing is, that feeling that something's
• 4:22
  wrong and I can't say what, that's the human dot. The fact that
• 4:26
  they can feel it means they know something their chapter doesn't yet contain.
• 4:31
  So the work we do together is basically using questions like bulldozers.
• 4:35
  Okay, but why does that matter to you? So what? Why is
• 4:38
  this important enough that you're writing about it in a book for strangers to read?
• 4:41
  Do you have a personal story that actually connects to this idea? Who,
• 4:45
  what, where, when, why? You know, the journalist questions, researcher questions.
• 4:49
  The questions you ask when you're trying to find out what's actually true
• 4:53
  underneath what somebody said. And what we're digging for is the human experience
• 4:57
  behind the message. Because sometimes people know what
• 5:00
  they want to say long before they understand why they want to say it's.
• 5:04
  And the why is where the pixel connects to the
• 5:08
  picture. It's the story behind the writing. And then
• 5:12
  there's this moment, and I love this moment, by the way. It never gets Old.
• 5:15
  Where something connects like a synapse, and your brain fires and
• 5:19
  they say something, and the client says, you know, the client says, oh, my God.
• 5:22
  Yeah, okay, now I get it. And that thing they've been trying to write around
• 5:26
  suddenly has a personal root. And once we have that, we can fill in
• 5:30
  the layers around the message that make it actually theirs and deeper and more specific
• 5:34
  and more true for them. You can't hand that entire
• 5:38
  process, that excavation process, to a machine.
• 5:41
  The machine doesn't understand all the little individual data points
• 5:45
  that I just described in that aha conversation. And this is
• 5:49
  what I mean when I say words are pixel, pixels of the human experience.
• 5:53
  Because when a book chapter feels technically accurate, but not
• 5:57
  actually mine, what's missing is that individual human experience underneath
• 6:02
  it. The chapter has the message, but not the root. The pixels
• 6:05
  are there, but they came from a pattern, maybe dictated by the peanut gallery
• 6:09
  of the outside world approval, but not from the author.
• 6:13
  AI is very good at producing that message without root
• 6:16
  language, without the experience that should be generating it. And it's fluent
• 6:20
  and organized, and it might even be moving in a generic way. But again,
• 6:24
  it's the trend line without the scatter plot, without all the individual data dots.
• 6:28
  And the reader can feel the difference. Difference. Research studies support this
• 6:33
  general idea that when humans write as themselves, you get more variety,
• 6:36
  which makes sense because every human writer brings their own scatter plot,
• 6:40
  their own individual dots, their own specific pixels of lived experience.
• 6:44
  AI just keeps pulling toward the averaged out trend. It keeps
• 6:48
  regressing to the mean. So I want to close with a story that I think
• 6:52
  is the most vivid example of what happens when the pixels are all there,
• 6:55
  but the human experience that should be generating them is not. In 2023,
• 7:00
  a New York attorney named Stephen Schwartz submitted a legal brief in a personal injury
• 7:04
  case against the Avianca Airlines. The brief cited six
• 7:08
  prior court decisions as precedent. All six cases,
• 7:12
  completely fabricated, made up. ChatGPT had invented them.
• 7:16
  Case names, judges, quotations, internal citations, all of it.
• 7:19
  But Schwartz submitted them to a federal judge.
• 7:23
  The cases had the structure of real legal cases. They had the right pixels,
• 7:27
  technically, more like the right trend line, but the underlying human experience.
• 7:30
  Experiences that should have been generating those words. Actual cases actually decided
• 7:34
  by actual judges involving actual people. None of those dots existed.
• 7:38
  The form was there. The shape on the. You know, the trend line was there,
• 7:42
  but the substance wasn't. And the federal judge noticed.
• 7:46
  Schwartz and his colleague were sanctioned $5,000, and the court issued
• 7:51
  letters to every judge who had been falsely named as the author of one of
• 7:54
  those invented decisions. The pixels looked right on
• 7:58
  some level. The picture was wrong. Your readers will notice that,
• 8:02
  too. Maybe not immediately, but maybe not consciously. But that feeling
• 8:06
  that something is technically fine and somehow empty, that's the missing scatter.
• 8:09
  That's the human experience that didn't make it into the words.
• 8:13
  So find a piece of your writing, something recent, or, you know, maybe something
• 8:17
  that AI helped you write, where something feels technically accurate, but not really
• 8:21
  yours. Ask yourself all those questions that I asked earlier.
• 8:25
  Why does it matter to me? So what? Why is this important enough to put
• 8:28
  in writing for strangers to read? Do I have a personal story that actually
• 8:32
  connects to this idea? Keep digging until you hit the individual
• 8:36
  experience, the personal, lived experience, under the message,
• 8:40
  the trend. Without the scatter dots. That's just a line. So once you find
• 8:44
  your specific dot, the writing will almost fix itself. It's like the dots
• 8:48
  in my survey, the actual scoliosis experiences, not just the trend of,
• 8:52
  you know, all of them together. Words are pixels of the human experience.
• 8:57
  Make sure that your pixels came
• 9:01
  from your experiences. Okay, that's it for
• 9:05
  today's episode of Human Print. And until next time, use the tool. If it
• 9:09
  helps, just don't hand it the pen.
• 9:20
  And that's it for this episode of Human Print. If you want a
• 9:24
  practical framework for protecting your voice amid the AI noise, please follow the show.
• 9:28
  And before you go, take a minute and ask yourself,
• 9:31
  as the AI noise grows louder and louder, which areas
• 9:35
  of yourself do you need to protect the most?
• 9:38
  I'll see you next time,