Latest / HumanPrint: How to Use AI Without Losing Yourself / Episode 17: Before AI / After AI

Transcript

Episode 17: Before AI / After AI · 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
• 0:17
  at one place where your authentic human signal is at risk for drift and
• 0:21
  one way to shore it up. Oh, and Human Print is brought
• 0:26
  to you by Abby AI Brand bank intelligence that helps
• 0:30
  authors, thought leaders and enterprise teams train AI
• 0:34
  on their voice and point of view so written content stays unmistakably
• 0:38
  human, consistent and on brand. Learn more@abbey.ink
• 0:44
  a b b I.ink like you write
• 0:48
  with. Let's begin.
• 0:52
  Hi everybody, and welcome back to Human Print.
• 0:55
  This is episode 17 and it's part two of
• 0:59
  my three part series, writing through the Noise.
• 1:02
  Protect your headspace, know your process, and keep creating.
• 1:06
  While everybody else argues about AI across
• 1:10
  these three episodes, I'm asking what gets access to your head, your creative
• 1:14
  process, and the work you haven't made yet? Last time in episode
• 1:18
  16, we dealt with your head. And today I want to get into the
• 1:22
  actual mechanics of how you write. Because before we have another gigantic
• 1:26
  philosophical argument about whether AI is good or bad for writers, it might
• 1:30
  be helpful to figure out what it has actually done to your writing.
• 1:33
  Not everybody else's, but yours. So consider this
• 1:38
  a little Writing in the Age of AI therapy session, if you will, like before
• 1:42
  AI and after AI. So just kind of all the questions in this episode
• 1:46
  just, you know, put in the context of before AI and after
• 1:50
  AI, what stayed? What changed? What got
• 1:54
  better? What did the technology make worse? And here's a big
• 1:58
  what have you changed because of the backlash around AI even
• 2:03
  when there was nothing wrong with that part of your process in the first place?
• 2:06
  There is a P.J. wodehouse story that I love for this. He was famously
• 2:11
  prolific and had a writing process that remained pretty
• 2:15
  consistent for decades. It was described in a 1971
• 2:19
  New Yorker profile that he wrote first drafts longhand
• 2:22
  in pencil, and then moved to a royal typewriter he'd been using since 1934,
• 2:26
  revising as he typed. At some point, he decided to try dictating
• 2:31
  a Jeeves and Wooster novel into a Dictaphone.
• 2:35
  He made it about one page, then he played it back, thought it sounded terrible,
• 2:39
  and immediately went back to his pencil and paper. Now that might be
• 2:44
  my ideal relationship with writing technology. Try the thing.
• 2:47
  See what happens to your work and if it helps, keep exploring it.
• 2:52
  But if it makes something worse, you now know something useful about your process.
• 2:56
  The arrival of a new tool doesn't create an obligation to use it in
• 3:00
  every available way. And deciding one particular use doesn't work for
• 3:04
  you. That does not require declaring the entire technology as evil.
• 3:08
  Wodehouse didn't need a global position on dictaphones. He just
• 3:12
  needed to know whether that Dictaphone helped him write a funny novel. Well, apparently it
• 3:16
  did not. That level of specificity that would help us enormously
• 3:20
  right now if we all applied that. So instead of asking whether AI
• 3:24
  makes writing better, ask what happened to your writing? And I'd start
• 3:28
  before generative AI even entered the picture at all.
• 3:32
  How did you used to get ideas? Where did they come from?
• 3:35
  Did you outline first or discover the idea while you were writing?
• 3:39
  How did you research what happened when you got stuck?
• 3:43
  What did revision look like for you? Who got to see your rough
• 3:47
  drafts? Which parts of the process did you love, and which
• 3:50
  ones did you mostly tolerate because they were necessary? How did you
• 3:55
  used to know the writing sounded like you, and what made something good
• 3:59
  quote unquote in your eyes before anybody was debating whether
• 4:03
  a machine might have touched it? These questions are your baseline.
• 4:06
  And that baseline is key right now, because one of the stranger side effects
• 4:11
  of the AI backlash is that perfectly normal human writing
• 4:14
  habits are being retroactively classified as suspicious.
• 4:19
  The EM dash has obviously become the all star celebrity
• 4:23
  example, and I say that as someone who was using the poor thing very happily
• 4:27
  long before it acquired this criminal punctuation record. But there
• 4:32
  are plenty of others, you know, there's the words, the bullet points, organized arguments,
• 4:35
  specific transitions, three part structures. Which is
• 4:39
  hilarious because I've always been obsessed with using groups of threes in my writing.
• 4:43
  I love threes. Also on the most wanted list, sentences that sound too
• 4:47
  polished. I've seen lists of AI tells that are so broad that I'm reasonably
• 4:52
  certain half of the English speaking population could be detained for questioning.
• 4:57
  So, okay, let's go back to your old writing. Seriously,
• 5:00
  pull something you wrote in maybe 2015, 2005,
• 5:04
  whenever, 1990s, whenever you have enough distance to know that
• 5:08
  generative AI absolutely was not involved. Which would be, I think,
• 5:12
  anything more than like a few years ago, what were you already doing?
• 5:16
  Which punctuation marks did you already like? What kinds
• 5:21
  of sentence structures already showed up for you? How did you transition between
• 5:25
  ideas? Which words kept appearing over and over again?
• 5:28
  Words, phrases, structures? How formal or informal were
• 5:32
  you? Maybe you discovered that something you've recently become self
• 5:37
  conscious about? Maybe it was yours all along. Then what?
• 5:40
  Are you going to stop sounding like yourself, because people might mistake your writing for
• 5:44
  a machine. That's a pretty weird outcome for a movement supposedly
• 5:49
  concerned with protecting human writing. And this is exactly why I keep
• 5:53
  going back to understanding your own process before defending it.
• 5:56
  I can make this comparison very clearly from ghostwriting,
• 6:00
  because my process was established long before generative AI ever existed.
• 6:04
  My version of ghostwriting has always been highly interview driven.
• 6:08
  I ask the client questions and then I listen for the real answers hiding
• 6:12
  underneath. Usually the first answers, those are the neat ones because
• 6:16
  that's often where the most interesting material is. It's the second or third or last
• 6:20
  answer. Sometimes there's a thing with ghost writing calls. I don't know why. This always
• 6:23
  happens the last 10 minutes after you. After I've said, all right, time to
• 6:27
  wrap it up. See you next time. All of a sudden, like, the best stuff
• 6:30
  starts spilling out. It's hilarious what happens in the last 10 minutes of a ghostwriting
• 6:34
  call. But I digress. So, yeah, I ask follow up questions
• 6:38
  and then I'll hear one phrase and I'll go back and be like, hey,
• 6:41
  back up. What did you mean by that? Tell me more. What happened next?
• 6:43
  Why did you make that decision? What were you thinking at the time? And eventually
• 6:48
  I have a pile of stories, ideas, observations, lessons, and memories
• 6:51
  and expertise that came from the client. So my job is to shape that material
• 6:55
  into this very smooth A to B journey for the reader.
• 6:58
  I'm doing the deep creative work. That's what the client hired me to do.
• 7:02
  But I have always drawn a hard line around putting ideas in
• 7:05
  someone else's mouth. So the point there is the boundaries that I was drawing around
• 7:09
  my process. If I have an idea that I think might supplement what
• 7:13
  a client is saying, I'll run it by them. I'll say, does this sound right
• 7:16
  to you? Does this fit what you believe? Am I understanding where you're going with
• 7:19
  this? I never assume that just because I'm thinking it, they're thinking
• 7:24
  it too. And again, that boundary existed way before ChatGPT,
• 7:28
  which is useful because it gives me something solid to compare against now.
• 7:31
  So I can ask, has AI changed my ghost writing
• 7:35
  process? Well, sure. It can help organize client thoughts, especially when
• 7:40
  there's a mountain of material to sort through. It can help with research,
• 7:42
  it can help me find connections in the information a little bit faster.
• 7:47
  But I'm still in the center of the process, like the circus ringmaster, deciding what
• 7:51
  happens next, deciding which stories deserve more space. What a
• 7:55
  reader needs to understand before moving on. What does not belong
• 7:59
  what needs another interview question? Or 5 or 10?
• 8:03
  Whether the structure is working and whether the finished material actually sounds like
• 8:07
  the human being whose name is on the COVID Now, that's my before and
• 8:11
  after. Some tools changed, but the principles governing what I'm
• 8:15
  willing to do with them stayed remarkably consistent.
• 8:18
  What does your version look like? Maybe AI legitimately
• 8:23
  improved part of your research process because you can get oriented to an unfamiliar subject
• 8:27
  faster and then go to the primary sources.
• 8:30
  Maybe you're better at organizing large amounts of your own material.
• 8:34
  Or maybe it helps you brainstorm questions that you hadn't even considered.
• 8:38
  Maybe you use AI to challenge an argument and discover where your reasoning
• 8:43
  is weak. Maybe it saves you hours doing something you never
• 8:47
  considered a meaningful part of authorship anyway. Well, good. But be
• 8:51
  able to say why it works for you. Then look at the other side without
• 8:54
  getting defensive about the technology. Where did it make you worse?
• 8:59
  Have you become more passive? Do you ask for an
• 9:02
  idea before giving your own brain time to come up with one?
• 9:06
  Are you accepting structures because they're reasonable or they sound
• 9:10
  reasonable rather than because they are the best structure for what you are
• 9:14
  writing? Do you spend so much time playing with prompts that the process has
• 9:18
  become longer than when you just sat down and wrote it on your own?
• 9:21
  Have you stopped practicing a skill that you actually care about maintaining?
• 9:26
  Does the writing sound less like you after the tool touches
• 9:30
  it? Those are good discoveries, and I'm very pro experiment.
• 9:34
  Even when the experiments fail. Sometimes, especially when experiments fail, you can learn a lot.
• 9:38
  So that Wodehouse dude, he got one page out of his Dictaphone and experiment
• 9:43
  before deciding, and he was like, nope, going back to the pencil. Yeah, that's information,
• 9:47
  that's feedback. And it's also consistent with something I used to
• 9:51
  tell writers constantly when they asked me about the quote, unquote,
• 9:54
  right way to write a book. Which software should I use?
• 9:58
  What time of day should I write? How many words should I write a day?
• 10:01
  Should I outline first? Now, I once tried that software Scrivener.
• 10:05
  Awesome software. A lot of people use it, a lot of my clients. And I
• 10:09
  spent what really felt, I think it was like an entire weekend just figuring out
• 10:12
  how to use it, all the bells and whistles. And I finally was just like,
• 10:14
  oh, I'm just going to go back to Microsoft Word and Google Docs, because those
• 10:17
  were working well for me. Me. And that's kind of my kind of litmus test
• 10:21
  over tool. Like is, if something's working well for me, why am I changing it?
• 10:24
  So that was less of a review of Scribner and more a decision about myself
• 10:28
  and my process. Use the methods and tools that help you
• 10:32
  sit down, do the work, and finish the thing.
• 10:36
  I don't see why Generative AI should be exempt from the same practical
• 10:40
  scrutiny as all those other tools. Another before and after question
• 10:44
  involves the actual material you're giving these tools, as I
• 10:48
  mentioned in the last episode, but I want to hit on again because it's such
• 10:51
  a big thing now. One conversation I've had for years with certain
• 10:54
  book clients, especially business ones, is about the desire to sound like they always
• 10:58
  had it all figured out. They were born knowing all the things about business and
• 11:02
  running successful companies or self help, motivation, productivity, all those
• 11:06
  things. They've always known it because they want to establish authority, which makes sense.
• 11:10
  So they start editing. All the uncertainty, all the vulnerability out of their story,
• 11:14
  their failures, their past failures disappear. So do the embarrassing
• 11:18
  parts and doubts. The years when they were absolutely winging it, they just
• 11:22
  all of a sudden they're gone. They never existed. And eventually
• 11:25
  I have to explain to the client that readers tend to trust you more when
• 11:29
  they understand how you arrived at the expertise. The story where
• 11:33
  you got something wrong and learned from it may carry more weight than three pages
• 11:37
  explaining the lesson after the fact. And this has become even more interesting
• 11:41
  with generative AI, because anyone can now ask a tool for
• 11:45
  10 pieces of advice about leadership, marketing, relationships, health,
• 11:48
  business, creativity, whatever. Information by
• 11:52
  itself has become incredibly easy to manufacture. Like I was talking
• 11:56
  about in the last episode, it's the abundant asset, your lived experience
• 12:00
  of that information, that's the scarce asset that's still yours.
• 12:04
  So when I'm ghost writing, I need the client to tell me everything.
• 12:07
  The weird thing is, when they want to censor the thought they believe isn't
• 12:11
  good enough, or they take out the story that they're afraid makes them look weak,
• 12:14
  or they nix the idea they haven't fully figured out yet. I mean, I'm a
• 12:18
  great interviewer, but I'm not psychic. I can only draw out what they're willing
• 12:22
  to give me. And I'm certainly not going to invent a convenient life experience because
• 12:27
  the chapter seems to need one. That same responsibility exists
• 12:31
  when you work with Gen AI. What material are
• 12:35
  you supplying? What do you know? What happened
• 12:39
  to you? What have you observed getting the theme here?
• 12:42
  Which opinions have you actually developed? Which questions are
• 12:46
  genuinely yours? Which part of the work would still exist if the
• 12:51
  tool disappeared tomorrow? There's one more category I want
• 12:55
  to look at, because it has Nothing to do with whether AI improved or damaged
• 12:59
  your writing. And that is what has the controversy
• 13:02
  changed for you as a writer. Maybe you
• 13:06
  genuinely reassessed an AI use after learning something
• 13:10
  new about it. That's fine. That's awesome. But maybe you're taking
• 13:14
  things that were working perfectly well and changing them because you are anticipating accusations.
• 13:19
  Maybe you stopped using your own natural punctuation style. Or you
• 13:23
  deliberately rough up your sentences to make them sound less smooth.
• 13:27
  You avoid words that you love, that you feel like belong to you. You put
• 13:31
  every draft through an AI detector, even though you know where the writing came from.
• 13:35
  You spend more mental energy wondering whether the text will look human than whether
• 13:39
  it will be useful to the reader, which should be the point of this whole
• 13:42
  thing. That's a process change, too, and I don't think we should ignore it just
• 13:47
  because the change came from outside the technology itself.
• 13:50
  So here's what I'd actually like you to do. Find something you wrote before
• 13:55
  generative AI was part of your writing life and put it beside something more recent.
• 13:58
  When it is a part of your life, then compare the processes,
• 14:02
  not just the words. Where did the ideas come from? What did you do
• 14:06
  yourself? Where did another human help? Where did a tool help?
• 14:10
  Which skills are stronger now? Which ones are getting a little rusty?
• 14:14
  What habits are recognizably yours across both pieces?
• 14:18
  And what have you started apologizing for or editing out
• 14:22
  simply because somebody has declared it suspicious?
• 14:25
  Now, this is why episode 13 of this show, this podcast,
• 14:30
  was about freedom of creative process. You can't govern a process, process you don't understand.
• 14:34
  So Next, in episode 18, we're taking the final step in my Writing through
• 14:38
  the Noise series. And we're going to look forward. Once you protect
• 14:43
  your headspace and understand your process, what gets access to the work
• 14:47
  you haven't made yet? Because there's a point
• 14:50
  where endlessly debating how everybody should create starts
• 14:54
  taking time away from actually creating.
• 15:06
  And that's it for this episode of Human Print. If you want
• 15:10
  a practical framework for protecting your voice amid the AI noise,
• 15:13
  please follow the show. And before you go, take a minute and ask yourself,
• 15:17
  as the AI noise grows louder and louder,
• 15:20
  which areas of yourself do you need to protect the most?
• 15:24
  I'll see you next time,