Latest / Spacious Success / 015 | The AI Addiction Problem: Why AI Tools Are So Hard to Put Down

Transcript

015 | The AI Addiction Problem: Why AI Tools Are So Hard to Put Down · Spacious Success

• 0:00
  Hi, it's Christy. Welcome back to the Spacious Success podcast. I want to start with a
• 0:12
  confession that maybe you have done as well. Perhaps you haven't said it out loud though.
• 0:20
  Last week I opened up Claude an AI tool. I used it to help me refine one paragraph for
• 0:27
  an article that I drafted. Yes, I used Claude as my cognitive sparring partner. It's certainly not
• 0:34
  my ghost writer. I looked up at the clock 45 minutes later, having gone through 11 prompts,
• 0:41
  three completely different structural approaches, and a D2 into researching a topic that I did not
• 0:48
  need to research at that point in time. I wasn't procrastinating. I wasn't distracted in that
• 0:56
  traditional sense that we might call distraction. I was completely focused, but I was focused on AI.
• 1:03
  I'd been sucked into the AI vortex. Does this ever happen to you? And when I finally closed down
• 1:09
  the Claude tool, I felt that particular kind of holiness that comes after too long on a screen.
• 1:15
  Do you know that same empty feeling if you've engaged in revenge bedtime procrastination,
• 1:21
  where you've promised yourself an early night, but you find that you've binge watched a Netflix
• 1:26
  series, and you've gone to bed three hours later than you'd intended. Maybe it's the same feeling
• 1:32
  when you realize that you've been scrolling on social media for an hour, when you simply went on
• 1:37
  there to find that one funny meme to send to your husband. It's that guilt and that panger of
• 1:45
  worry and remorse that tells you that something in your brain just got used in a way that it wasn't
• 1:51
  quite designed for. Today, I want to talk about why that keeps happening, not as a cautionary
• 1:58
  tale, and not because I think AI is bad. I don't. I use it daily. I think it's genuinely powerful,
• 2:04
  but I think we need to understand what it's doing to our brains, specifically why it's so hard
• 2:12
  for us to stop using it, because that understanding, I believe, is what will give us back control
• 2:19
  and help us to leverage the benefits that AI offers us.
• 2:23
  Hi, I'm Dr. Christie Goodwin, a neuroperformance scientist, speaker and executive coach who learned
• 2:29
  the hard way that chasing success can come at a cost. After burnout landed me in hospital,
• 2:36
  I made it my mission to help ambitious women who are done trading their health and happiness
• 2:43
  for high performance. On spacious success, I share science-backed strategies and real stories
• 2:51
  to help you flourish as a female at work and beyond so you can stop paying the success tax.
• 3:03
  Recently, I was running a workshop for a leadership team. This was a team of highly
• 3:08
  capable people. There were a mix of directors, general managers, most of them enthusiastic
• 3:14
  early adopters of AI tools. I will admit there were a few people who were reluctant users.
• 3:20
  About halfway through the session, I asked everybody to close their laptops and work through
• 3:25
  a strategic problem using what was in the room. No AI, no Google, just the collective expertise
• 3:34
  of the 12 experienced leaders who were sitting in that room. And I will admit, the discomfort
• 3:40
  was palpable. It was immediate, it was visible. Not because they couldn't do it, they absolutely
• 3:46
  could, but because their first instinct was to reach for AI and that was so automatic, so deeply
• 3:52
  ingrained already, that sitting with the uncertainty of not immediately having an answer
• 3:59
  felt for that group genuinely uncomfortable, discombobulating. One of the directors said to me
• 4:05
  in the afternoon tea break, I kept thinking, why are we doing it the hard way? And then I realized
• 4:11
  that I couldn't remember what the hard way felt like before AI existed. And that really worried me.
• 4:20
  That is the thing that I want us to examine in today's episode. This is not a conversation about
• 4:26
  whether AI is useful, it is. But why does it have such a strong pull? Why do we reach for it,
• 4:32
  even before we've had an opportunity to give ourselves time to think? Why does one prompt become
• 4:37
  11 prompts? Why is closing down Claude or co-work or co-pilot like leaving, feel like leaving something
• 4:46
  unfinished? I'm going to answer a DM that I received on Instagram recently and I'm going to read
• 4:51
  it out. It said, why was I using Claude at 2am when I started using it at 9.30 pm just for one
• 4:59
  task? I'm a smart woman with a career but I got drawn in and spat out four and a half hours later
• 5:06
  like a pubescent boy addicted to his gaming console. You see this is not a character flaw,
• 5:14
  it is neuroscience. AI is by design neurobiologically compelling and once you understand the
• 5:20
  mechanisms then you can start to work with them rather than them be unwittingly driven by them.
• 5:28
  Let me walk you through the five reasons that I think AI is so appealing, why is it so hard to
• 5:35
  put it down? Each one is rooted in how your brain actually works and when you stack all five
• 5:42
  together you start to understand why this is not a willpower problem, you are not flawed,
• 5:47
  you are not weak. This is technology being engineered to be irresistible. I'm going to
• 5:55
  deliberately avoid using the word addicted and here's why. The moment we call it an addiction,
• 6:01
  I've got an AI addiction as I've heard many people confessing, we hand away our agency. There's
• 6:06
  no clinical definition yet of an AI addiction and more importantly there are actually very real
• 6:12
  things that you can do to make this technology less sticky, more friction between you and the
• 6:18
  technology. This isn't about diagnosing you, it's just simply about equipping you with knowledge.
• 6:25
  The first thing is that dopamine loops and instant feedback make AI so appealing.
• 6:31
  Every time you submit a prompt and you receive hopefully a high quality useful response,
• 6:37
  your brain releases a little hit of the neurotransmitter dopamine. This is the same neurochemical reward
• 6:43
  loop that often drives every other digital compulsion, social media, email, gambling, gaming,
• 6:50
  online shopping, but AI I think has a particular edge. The reward is not random. It's almost always
• 7:00
  good. It's usually useful, relevant, and that consistent high quality feedback is neurologically
• 7:07
  more compelling than the variable rewards of scrolling because your brain quickly learns
• 7:13
  that your effort to reward ratio is rather good. In fact, I might even say it's usually excellent.
• 7:19
  Why would you wrestle with a problem yourself when the dopamine hit of a very polished AI
• 7:25
  response or output is only three seconds away? The second reason is what I call intermittent
• 7:31
  variable rewards, and here's where I think it gets a little bit more sophisticated.
• 7:36
  While I just said the average AI output is usually pretty good, occasionally the output will
• 7:43
  be extraordinary, but sometimes it's downright awful or worse still is when it makes things up.
• 7:50
  Yes, there is a real AI phenomenon called AI hallucination, but that's a topic for another day.
• 7:57
  When AI's output is superb or genuinely surprises you, that's when it's captivating.
• 8:04
  If it's better than you're expected, if it's better than what you could produce, it's more creative,
• 8:09
  it's more insightful, it's more useful than you thought was even incomprehensible.
• 8:15
  Neuroscience tells us that that intermittent variable reward ratio are sometimes
• 8:21
  more captivating. They are a powerful driver of compulsive behaviour. You see this is the same mechanism
• 8:31
  that makes poker machines so hard to walk away from. You don't know when you're going to hit
• 8:36
  the next jackpot from the prompt that you put in. Will it be ordinary or will it be brilliant?
• 8:42
  So you keep going, one more version, one more angle just to see what the output is.
• 8:48
  The third reason is novelty. Our brain has a dedicated novelty detection system,
• 8:55
  new information, new perspectives, new findings, new framings of a familiar problem.
• 9:01
  These trigger that dopamine that I just mentioned independently of whether the content is actually
• 9:07
  useful, AI generates endless novelty. Every prompt or next prompt can produce output that you've
• 9:15
  never even seen or contemplated before. Every iteration offers a new angle, a new structure,
• 9:21
  a new way of seeing or considering something, and your brain reads this as valuable. It's stimulation,
• 9:28
  even when the 10th version really isn't materiality, any better than the third one. The novelty
• 9:35
  alone keeps that loop running. The fourth reason is one that I find fascinating. It's self-determination
• 9:42
  theory, and this one's not as obvious, but I think it's important. Self-determination theory was
• 9:48
  developed a while ago by two researchers, Ryan and Dessie, and they identified three psychological
• 9:55
  needs that drive motivation in humans, competence, connection, and autonomy or I call a control.
• 10:03
  An AI satisfies all three simultaneously and with very low friction.
• 10:09
  Competence. Every interaction you make with AI makes you feel capable,
• 10:14
  informed, productive. The AI is constantly pumping up your tires and affirming your direction.
• 10:20
  Connection. There's a genuine sense of being understood of having a responsive interaction
• 10:27
  that adapts to you. People are legitimately forming para-social relationships with AI tools,
• 10:34
  hence why panic sets in when you've used up all of your claw tokens for the day.
• 10:38
  And that last one control or autonomy. We are in control of every prompt, every direction,
• 10:45
  every output. You can change your mind instantly, and AI will bend to you.
• 10:51
  No other tool in your workday consistently meets those three psychological drivers all at once,
• 10:58
  which means that it's not just satisfying your cognitive needs. At a deep psychological level,
• 11:04
  it is meeting the most basic psychological drivers we have as humans, connection,
• 11:10
  competence, and control. And the fifth one is that we really need to exert minimal cognitive
• 11:15
  effort, but with maximum output. This one is where the long-term risk lives.
• 11:22
  AI dramatically reduces the cognitive effort that we have to use to produce something that looks
• 11:29
  good and feels high quality. A paragraph that would have taken 20 minutes of focus drafting,
• 11:35
  changes, and edits now takes 30 seconds of prompting. Recently, I had clawed,
• 11:41
  helped me to co-create a digital assessment tool that would have previously taken me days to
• 11:49
  develop, and I would have had to pay a developer a significant amount to build it. This whole project
• 11:55
  took less than a minute to create a finished product. A strategic framework that would have
• 12:01
  required an hour of structured thinking can appear on your screen in two minutes,
• 12:05
  and your brain registers this output as a success while bypassing the
• 12:12
  cognitive rigor, the cognitive effort that you would have had to exert to produce that.
• 12:17
  And over time, this is what researchers are calling cognitive offloading.
• 12:23
  AI or brain atrophy. We're offloading tasks that are genuinely useful.
• 12:29
  We should be offloading our effortful thinking, our expertise that sharpens our judgment.
• 12:38
  Research from ActiveTrack in 2026 that tracked 164,000 workers found that the time that we're
• 12:45
  spending in deep-focused work has dropped by 9% following the widespread adoption of AI,
• 12:52
  even though our output measures are increasing. You see, this tells us that we're working faster,
• 12:58
  but I think we may be thinking less deeply. Now, stack all five of these mechanisms together,
• 13:05
  and you have a tool, an AI tool that's delivering dopamine hits, it's giving you unpredictable
• 13:10
  rewards, it's offering you constant novelty, and it's the psychological needs satisfaction and
• 13:17
  effortless output all at once. Now, of course, it's hard to put down. It was built this way.
• 13:25
  So when you clock up another hour on co-pilot, and you wonder where the time went,
• 13:30
  or why you're still there finessing the output three hours later, please know it's not you.
• 13:37
  The question is not whether you should use AI. I think that's a redundant conversation.
• 13:41
  The question is whether you are using it with intention, or whether it is using you.
• 13:47
  I want to share three strategies for using AI with intention, rather than with what I'd call
• 13:53
  compulsion, and I didn't say addiction. The first one is to set a prompt limit before you open it
• 13:59
  or a time limit. Decide in advance how many prompts or how much time you think this task will require,
• 14:06
  one for a quick refrain, three for a first draft, five for a complex research task. Write it down,
• 14:13
  set a timer. This interrupts the state of insufficiency, which is that feeling that we never quite
• 14:20
  feel being finished by giving you a defined endpoint. When you hit your limit, the number of prompts
• 14:27
  or your time limit, close down the AI tool. The task is done, and this sounds almost
• 14:32
  comically too simple, but believe me, it works precisely because it externalizes the stopping
• 14:40
  decision, removing it from that dopamine loop entirely. The second thing is what I call a think
• 14:46
  first protocol. Before you open any AI tool for a substantive task, spend just a moment,
• 14:53
  five minutes working on the problem independently and in an analog way. I often say we need AI
• 14:59
  to be your sparring partner. It should not be your ghost writer. Write down your initial thinking,
• 15:05
  create a concept map, do what I call a brain sneeze, sketch a rough structure or an outline,
• 15:12
  name the key questions that you want to address. This does two things. It means that it's your
• 15:17
  perspective and expertise that are shaping the prompt rather than the AI shaping your thinking
• 15:23
  from the outset. It keeps your neural pathways for independent problem solving active.
• 15:31
  That cognitive capacity that you protect by thinking first is exactly the capacity that makes
• 15:38
  your AI assisted output better than anyone else's. It stops you from producing AI junk.
• 15:45
  Your expertise, your experience are what the AI cannot replicate. Use them first.
• 15:52
  And the third one is to designate AI-thinking, AI-free thinking time.
• 15:57
  We need to categorize time so that we can have that initial creative phase,
• 16:03
  the first draft of communications that will carry your voice. That strategic thinking time that
• 16:09
  draws on institutional knowledge that only you hold that hasn't been put into the knowledge
• 16:14
  base of the tool. Not because AI can't contribute to these things, but because the process of
• 16:19
  mentally wrestling with ideas is where your expertise will sharpen. It is where your
• 16:25
  perspective will develop and it's your human potential. Protect those spaces like you would
• 16:31
  protect anything that is finite and valuable. Because your capacity for original thought,
• 16:38
  for problem solving and for ideation is exactly that. This week's power up task is to set your
• 16:47
  time or your prompt limit. Choose one task where you would have normally quickly reverted to AI
• 16:54
  and make a promise that before you open it, write down the number of prompts or the amount
• 16:59
  of time that you want to spend using it and also acknowledge what does done look like.
• 17:04
  Then hold that line. Notice what happens when you hit your limit. When your clock reminds you that
• 17:11
  your time is up or you've hit your five prompts. That sense that it could be better, that's that
• 17:17
  state of insufficiency doing its thing. It's exactly what happens when one YouTube clip rolls into
• 17:22
  the next. This happens on AI, so give the AI tool your terms rather than strictly adhering to its
• 17:30
  terms. My power up performance keynote explores how to maintain peak cognitive performance
• 17:37
  in an AI enhanced world and it's what leaders and organisations need to understand to protect the
• 17:44
  human capabilities that AI cannot and will not replicate. If you're interested in bringing this
• 17:51
  conversation to your next event, please reach out at drchristiglwant.com. I'm drchristiglwant.
• 17:58
  Thank you for listening to today's episode of spacious success.