Latest / How I Work / This is what AI is actually doing to your brain, with Gabriella Rosen Kellerman
Transcript
- 0:01Speaker 1: It's late afternoon and you've got three AI tools running
- 0:04Speaker 1: at once, browser tabs, multiplying, and your brain just stops working.
- 0:10Speaker 1: You're not burnt out, but something is definitely wrong. Gabriella
- 0:16Speaker 1: Rosen Callaman is a leader at Boston Consulting Group, and
- 0:20Speaker 1: she actually decided to study that feeling and give it
- 0:24Speaker 1: a name. AI brain fry. Brain fry is not the
- 0:29Speaker 1: same as burnout. It doesn't follow the same rules, and
- 0:33Speaker 1: the things that organizations are doing to fix it, well,
- 0:36Speaker 1: some are actually making it worse. In this conversation, Gabriella
- 0:41Speaker 1: and I get into who is getting hit hardest, why
- 0:45Speaker 1: using more than three AI tools at once is where
- 0:48Speaker 1: productivity falls off a cliff, and what managers are doing,
- 0:52Speaker 1: often unknowingly that is adding fifteen percent more mental fatigue
- 0:56Speaker 1: to their teams. This episode really changed how I think
- 1:01Speaker 1: about my own AI use, and I hope it does
- 1:04Speaker 1: the same for you. Also, just a quick heads up,
- 1:07Speaker 1: there's a little bit of background noise at the very
- 1:08Speaker 1: start of this episode. Record it early so it clears
- 1:11Speaker 1: up quickly to stick with us.
- 1:19Speaker 2: Welcome to How I Work, a show about habits, rituals,
- 1:22Speaker 2: and strategies for optimizing your day. I'm your host, doctor
- 1:26Speaker 2: Amantha imber.
- 1:32Speaker 1: Gabriella. Your article all about brain fry for HBr went
- 1:37Speaker 1: completely nuts? Can you tell me what is brain fry?
- 1:41Speaker 1: For those that have no idea.
- 1:43Speaker 3: I'll start by giving you the definition that we used
- 1:46Speaker 3: in the study, and then I'll break it down kind
- 1:49Speaker 3: of in layman's terms. So we define it as mental
- 1:54Speaker 3: fatigue caused by excessive use or oversight of AI tools
- 1:59Speaker 3: beyond what one's cognitive capacity. And think of that as
- 2:03Speaker 3: like a strain of your brain from over using AI
- 2:07Speaker 3: in mentally intense ways.
- 2:10Speaker 1: And how is that different from just burnout?
- 2:13Speaker 3: Let me start by actually explaining how we came to
- 2:16Speaker 3: do this study to begin with, which was we were
- 2:19Speaker 3: really interested in what is the relationship of burnout to
- 2:23Speaker 3: the use of AI tools at work? And it turns
- 2:26Speaker 3: out in the literature that it's very unclear. So some
- 2:32Speaker 3: studies suggest that when you use a lot of AI
- 2:34Speaker 3: tools at work, it makes the workload lighter, it gives
- 2:38Speaker 3: you more free time, and it decreases burnout. And then
- 2:42Speaker 3: other studies suggest that when people are using more AI
- 2:45Speaker 3: tools at work for a variety of unclear reasons and mechanisms,
- 2:49Speaker 3: that may be increasing burnout, and it even happens to
- 2:52Speaker 3: be the case that there's a whole literature studying radiologists
- 2:57Speaker 3: and whether radiologists in part particular have increased or decreased
- 3:02Speaker 3: burnout because of the use of AI tools. So we
- 3:05Speaker 3: were very interested to study it, and in particular, as
- 3:10Speaker 3: you may be aware, there is more and more usage
- 3:13Speaker 3: of multi agents at work, meaning you're not just using
- 3:16Speaker 3: one AI tool at once, you're using multiple agents at once.
- 3:19Speaker 3: Sometimes the agents are working with each other. And what
- 3:23Speaker 3: we found was that indeed there's this way of using
- 3:27Speaker 3: AI very intensively. We're really trying to keep up with
- 3:31Speaker 3: these tools and use a lot of them at once,
- 3:33Speaker 3: and it's very cognitively intense, and it produces this experience
- 3:38Speaker 3: that we call AI brain fry that does not have
- 3:41Speaker 3: a relationship with burnout, doesn't increase burnout, doesn't decrease burnout.
- 3:46Speaker 3: And then we did see that totally different type of
- 3:49Speaker 3: AI use, which is using a I to replace repetitive
- 3:53Speaker 3: tasks at work, decreases burnout. But we hopefully also started
- 3:58Speaker 3: to advance the conversation about the extent to which burnout
- 4:01Speaker 3: as a construct is no longer sufficient to help us
- 4:06Speaker 3: understand the domains of strain that will be raised in
- 4:11Speaker 3: the new era of AI work.
- 4:12Speaker 1: I find that's so interesting the distinction. Now in your research,
- 4:17Speaker 1: you look at the different types of roles or different
- 4:19Speaker 1: functional areas within an organization, and interestingly, marketing people had
- 4:25Speaker 1: the highest rates of brain for it. I think it
- 4:27Speaker 1: was twenty six percent. What's going on there, Like, how
- 4:32Speaker 1: are marketing people different from say, finance people.
- 4:35Speaker 3: Yeah, so it's a great question. The first thing I'll
- 4:39Speaker 3: say is that it's possible that marketing is on the
- 4:44Speaker 3: higher end, but maybe there will be other roles that
- 4:47Speaker 3: spake higher in future studies. What we did see in
- 4:50Speaker 3: general is that roles that are more operationally heavy, roles
- 4:53Speaker 3: that are more technically heavy, seem to be spiking higher
- 4:56Speaker 3: than other roles. And indeed, in marketing there is a
- 5:01Speaker 3: lot of operational lenses that are brought to the use
- 5:04Speaker 3: of AI, things like monitoring content, looking at optimizing content particularly,
- 5:10Speaker 3: and like search engine optimization, you know, question and answer optimization.
- 5:16Speaker 3: Marketing is one of the domains where a lot of
- 5:18Speaker 3: the roles have been most disrupted by AI in terms
- 5:22Speaker 3: of the specific skills and responsibilities. So in a way,
- 5:26Speaker 3: it wasn't surprising to us to see that. I just
- 5:29Speaker 3: want to be cautious in how we extrapolate from that
- 5:33Speaker 3: to the level of certainty that we can get to.
- 5:36Speaker 1: So your research found that it's about AI oversight specifically
- 5:40Speaker 1: as opposed to using AI, say for automating or you know,
- 5:44Speaker 1: offloading repetitive tasks. Can you explain what AI oversight is
- 5:51Speaker 1: specifically like instead of day to day layman's terms.
- 5:55Speaker 3: Yeah, so it means quite a lot of things. But
- 5:58Speaker 3: if you've ever inn with an AI agent, you'll notice
- 6:02Speaker 3: that you're going to give it a prompt, it's going
- 6:05Speaker 3: to give you something back. You're going to respond to that.
- 6:08Speaker 3: You are directing the agent at work. That agent might
- 6:13Speaker 3: be responsible for actually delivering value, let's say some analysis
- 6:18Speaker 3: or delivering code. Right, So, to the extent that you
- 6:22Speaker 3: are then responsible for the output of that agent or
- 6:26Speaker 3: that tool, you are overseeing it, and your domain of
- 6:31Speaker 3: responsibility now extends to that agent. In a way, you
- 6:36Speaker 3: could say that you are managing that agent as a manager.
- 6:39Speaker 3: It's different type of management, but it is true that
- 6:42Speaker 3: your sphere of accountability has increased. The activities include decision making,
- 6:49Speaker 3: error correction, feedback, prompts, monitoring all kinds of interactions that
- 6:57Speaker 3: are required in order to oversee that agent.
- 7:01Speaker 1: And so in terms of the language around agents, so
- 7:04Speaker 1: that could just be Chetchjpat or co pilot or Claude
- 7:09Speaker 1: or Gemini as opposed to true a gentiki.
- 7:13Speaker 3: Those are agents. So many of those it could be
- 7:17Speaker 3: those things. It could be agents that are running other
- 7:20Speaker 3: agents and you are responding to sort of the orchestrator layer.
- 7:25Speaker 3: Sometimes some of the folks in the study talked about
- 7:28Speaker 3: agents that are running longer processes and then sort of
- 7:31Speaker 3: feeding back to them and they then have to go
- 7:34Speaker 3: through and figure out what may have gone wrong at
- 7:37Speaker 3: various points in the process, so unwinding it versus the
- 7:41Speaker 3: more conversational agents where it's more one step at a time.
- 7:46Speaker 3: But yes, all of those things could be under the
- 7:49Speaker 3: umbrella of oversight and monitoring.
- 7:51Speaker 1: And it was interesting because you looked at how many
- 7:55Speaker 1: AI tools increase productivity, and then there was this productivity
- 8:01Speaker 1: cleif with productivity tanked after I think it was about
- 8:05Speaker 1: three AI tools. Why is two or three AI tools
- 8:10Speaker 1: the optimal number? And what happens after that?
- 8:13Speaker 3: For those who haven't seen the study, what we did
- 8:15Speaker 3: was we asked people how many tools are you using
- 8:18Speaker 3: at once, and it was open ended. Many people said one,
- 8:22Speaker 3: many people said two, one person said twenty five, and
- 8:26Speaker 3: then everything in between. And then we you know, we
- 8:30Speaker 3: asked all kinds of other questions, including things like how
- 8:33Speaker 3: much has AI increased your productivity at work? It happened
- 8:37Speaker 3: to be the case there was a really interesting relationship
- 8:39Speaker 3: between those two particular items, such that this self reported
- 8:45Speaker 3: productivity increased through AI increased with multiple use of tools
- 8:50Speaker 3: from one to two, from two to three, but then
- 8:54Speaker 3: beyond three that self reported increased decreased, the point being
- 8:59Speaker 3: there is diminishing returns beyond about two to three tools.
- 9:04Speaker 3: It may vary by person. By the way, our brains
- 9:06Speaker 3: are not all the same. I think it's important also
- 9:09Speaker 3: to not like this is timestamped to January of twenty
- 9:12Speaker 3: twenty six and to the generation of tools that we
- 9:15Speaker 3: have right now. Oftentimes, what is a mean to work
- 9:18Speaker 3: with these tools all at once is the tools have
- 9:21Speaker 3: processing times, and so maybe it's the case that in
- 9:25Speaker 3: the windows, where one tool is processing, you can go
- 9:28Speaker 3: to another tool and turn your attention and then it's
- 9:30Speaker 3: sort of like a triangle of rotation between the tools,
- 9:34Speaker 3: and if there's faster processing times, maybe you can't do that.
- 9:38Speaker 3: Another thing that didn't make the cut of the study
- 9:41Speaker 3: of what we were able to fit into the publication
- 9:43Speaker 3: that I'd love to share is that we saw a
- 9:46Speaker 3: similar pattern of the peak at three for the sophistication
- 9:52Speaker 3: of the users of those tools. So we have at
- 9:56Speaker 3: BCG a categorization of AI sophistication that's really just about
- 10:01Speaker 3: how long you've been working with AI tools, how well
- 10:05Speaker 3: you know how to work with orchestrateor agents and autonomous layers,
- 10:09Speaker 3: and it helps us see how do we learn to
- 10:13Speaker 3: get better with the tools over time and therefore say
- 10:17Speaker 3: things about like what's the limitation of someone's experienced curve
- 10:21Speaker 3: or learning curve versus the tools, let's say, or some
- 10:24Speaker 3: interaction between human and AI that's more fundamental. And what
- 10:28Speaker 3: we see is that it seems to be that the
- 10:31Speaker 3: most sophisticated users tend to prefer to be around three.
- 10:35Speaker 3: So there are the folks who are hanging out, you know,
- 10:39Speaker 3: maybe at the twenty five that I mentioned are probably
- 10:42Speaker 3: not the most sophisticated users are probably folks who are
- 10:45Speaker 3: still a little bit earlier in their learning curve. Obviously,
- 10:48Speaker 3: they're not the earliest, but there's some wisdom that's coming
- 10:52Speaker 3: into play around. It's not about like maxing out the
- 10:56Speaker 3: number of tools you can use at once. It's really
- 10:59Speaker 3: finding that sweet spot of productivity where it feels good
- 11:02Speaker 3: to your brain. You have that feeling of mastery and
- 11:05Speaker 3: yet you're getting that boost from the tools with the productivity.
- 11:09Speaker 1: I think about my own use of AI, and it's
- 11:12Speaker 1: really interesting, you know that you point out, Okay, this
- 11:15Speaker 1: study is time dated to January twenty twenty six. And
- 11:19Speaker 1: depending on what I'm asking my AI tools to do,
- 11:23Speaker 1: and I'm generally more of a claud user, although I'll
- 11:26Speaker 1: sometimes go into chat TOPT for certain things. Is that
- 11:30Speaker 1: you know, processing times do vary. Like if I'm asking
- 11:34Speaker 1: the AI tool to do some deep research on a
- 11:37Speaker 1: topic and meanwhile I might have another tab or task
- 11:42Speaker 1: open that is, perhaps I'm working with it to edit
- 11:46Speaker 1: some content. Then I know for me, and I mean,
- 11:50Speaker 1: I know how bad multitasking and context switching is for
- 11:54Speaker 1: my brain as a psychologist, but I find myself doing
- 11:57Speaker 1: it quite a bit. When I'm using the AI tools
- 12:01Speaker 1: to help augment or save time on whatever it is
- 12:04Speaker 1: that I'm doing, I'm curious is that something you can
- 12:09Speaker 1: relate to in your gabriella And I'm wondering what behaviors
- 12:13Speaker 1: have you changed around how you're interacting with AI since
- 12:16Speaker 1: doing this study.
- 12:17Speaker 3: One fascinating thing in the in the qualitative descriptions of
- 12:22Speaker 3: brain fry, there was a lot of task switching that
- 12:25Speaker 3: was mentioned as we ask people what does brain fry
- 12:28Speaker 3: feel like and when have you experienced it? A lot
- 12:32Speaker 3: of discussion of you know, I'm bouncing between browser tabs
- 12:36Speaker 3: and what it feels like to go from tool to tool.
- 12:39Speaker 3: I think that element of AI is very much connected
- 12:43Speaker 3: to the broader issue of digital overload, which you know,
- 12:46Speaker 3: we've been talking about for a couple of decades now.
- 12:51Speaker 3: So for me taking breaks having like true no digital
- 12:57Speaker 3: slots and I do at least twenty for hours breaks
- 13:01Speaker 3: from the digital once a week. I think having breaks
- 13:05Speaker 3: at work where I do phone calls instead of zoom
- 13:08Speaker 3: calls and I'm away from my computer and just talking
- 13:12Speaker 3: and doing the audio is really really helpful and restorative.
- 13:16Speaker 3: I think right now where there is like a feeling
- 13:20Speaker 3: of excitement and such deep engagement with the AI tools
- 13:24Speaker 3: and a feeling of like, look at this magic that's
- 13:27Speaker 3: very hard to pull away from. So the more we
- 13:31Speaker 3: can be aware of the fact that just being engaged
- 13:35Speaker 3: in this hyper busy way with our computers, with any
- 13:39Speaker 3: kind of digital tool, there are brain networks we can't
- 13:42Speaker 3: access when we operate that way, and knowing to take
- 13:45Speaker 3: those breaks and take those digital detox moments, you will
- 13:50Speaker 3: feel different, right, It feels different in your brain and
- 13:53Speaker 3: your body when you step away from these devices phones too,
- 13:57Speaker 3: like that the audio is fine, but the internet, the email,
- 14:01Speaker 3: that's all part of the same thing is a very
- 14:04Speaker 3: important part of my routine as well.
- 14:07Speaker 1: Now you give some recommendations for what leaders and managers
- 14:12Speaker 1: can do for their teams that are presumably using AI
- 14:17Speaker 1: quite frequently. What are sort of the top ways that
- 14:22Speaker 1: managers can improve this brain for our situation?
- 14:26Speaker 3: Yeah, thank you for asking. One of the most useful
- 14:31Speaker 3: parts of this study was we were able to demonstrate
- 14:35Speaker 3: these strong predictive relationships between manager behaviors, team practices and
- 14:42Speaker 3: organizational factors and mental fatigue. And these factors were all
- 14:47Speaker 3: AI related. So at the manager level, the two most
- 14:51Speaker 3: important factors that we found on the one hand, when
- 14:55Speaker 3: managers spend more time answering their employee questions around AI
- 15:02Speaker 3: that predicted fifteen percent less mental fatigue for the members
- 15:07Speaker 3: of their team, which is a huge number. And the
- 15:12Speaker 3: way I tell myself the story of that is that
- 15:16Speaker 3: this is a manager who's leaning into the inner personal time,
- 15:20Speaker 3: the human relationships, the care and the support for the team.
- 15:25Speaker 3: You know, in the best case scenario, these tools are
- 15:28Speaker 3: giving us more time for that. They're giving us more
- 15:31Speaker 3: time to feel connected, to feel that sense of belonging
- 15:35Speaker 3: in community, and we're experiencing them together as a team.
- 15:40Speaker 3: When the manager is showing up that way, that is
- 15:43Speaker 3: having a sense of you know, slve, it's really showing
- 15:48Speaker 3: this decreased effect on the mental fatigue of the team.
- 15:52Speaker 3: And then the inverse of that came through as well,
- 15:55Speaker 3: in the sense that there was a five percent increase
- 15:58Speaker 3: in mental fatigue of the team. When the manager is
- 16:02Speaker 3: setting an expectation that the employees should just be learning
- 16:06Speaker 3: how to use the tools on their own. So when
- 16:09Speaker 3: this is framed as like a solo mission, go out there,
- 16:13Speaker 3: do this on your own, it's not about us doing
- 16:16Speaker 3: this together, there is an increase in the mental fatigue
- 16:19Speaker 3: that the team members will experience.
- 16:22Speaker 1: I find that so interesting. Now, how about some of
- 16:25Speaker 1: the company messages that I'm certainly hearing, Like a lot
- 16:28Speaker 1: of organizations sprout like you know, AI is going to
- 16:31Speaker 1: make you more productive? Like what do those messages do
- 16:36Speaker 1: to brain fry burn out, having higher intentions to leave
- 16:41Speaker 1: your company? Like what messages should leaders be sending about AI?
- 16:46Speaker 3: The message that had the strongest impact in terms of
- 16:50Speaker 3: decreasing burnout in particular and mental fatigue was sending a
- 16:56Speaker 3: message from that the company cares about work life balance.
- 17:00Speaker 3: So this was a case where it actually is not
- 17:03Speaker 3: a message about AI. It's really a cultural message of
- 17:07Speaker 3: the value that's placed on having a workplace and an
- 17:12Speaker 3: employee base that is fulfilled, that has a sense of
- 17:17Speaker 3: that balance between who they are as a person and
- 17:20Speaker 3: who they are as a professional. Having that come through
- 17:23Speaker 3: as a strong positive pays off dividends in terms of
- 17:28Speaker 3: the pros and the engagement and the productivity that we're
- 17:31Speaker 3: seeing in the data. In terms of the AI story,
- 17:35Speaker 3: what we did see in the study, which is important
- 17:38Speaker 3: to know and important to take forward, employees who felt
- 17:41Speaker 3: the sense that their workload was going to increase dramatically
- 17:45Speaker 3: because of AI were more vulnerable to this mental fatigue,
- 17:50Speaker 3: and there could be lots of reasons for that, but
- 17:54Speaker 3: the goal shouldn't be to tell the story that your
- 17:56Speaker 3: work's going to get heavier because of AI. And I
- 18:00Speaker 3: don't think organizations are intentionally telling a story like that, right.
- 18:04Speaker 3: It's more about how do we emphasize the positives that
- 18:08Speaker 3: will come for AI. How do we emphasize the sense
- 18:11Speaker 3: of there is a lot of creativity that can open up.
- 18:16Speaker 3: There's a lot of new domains of building that you
- 18:19Speaker 3: can do as an individual. There's a lot of lift
- 18:22Speaker 3: that you can get through this that you things you
- 18:25Speaker 3: have not been able to do before. Lean into those
- 18:29Speaker 3: domains of it, and really watch out for unintentional undertones
- 18:34Speaker 3: of this is going to lead to that increased work
- 18:39Speaker 3: which people, when they hear, will be more disposed to
- 18:42Speaker 3: that mental fatigue.
- 18:43Speaker 1: I think that's so interesting because something that like I
- 18:47Speaker 1: certainly hear the wellbeing message from ladies and I even remember,
- 18:51Speaker 1: I mean, this is going back a couple of years
- 18:52Speaker 1: ago when AI was I guess you know, when it
- 18:55Speaker 1: was pretty early in the journey for a lot of businesses,
- 18:58Speaker 1: and I remember speaking to achieve people officer at a
- 19:01Speaker 1: big global firm, and they'd been copying a lot of
- 19:05Speaker 1: flak in the media for having very high workloads and
- 19:09Speaker 1: having very high levels of burnout. And I remember talking
- 19:14Speaker 1: to her about her strategy and she said, well, AI
- 19:16Speaker 1: will solve that because we'll implement it, we'll train people,
- 19:20Speaker 1: and then they'll be able to do essentially I'm paraphrasing
- 19:23Speaker 1: a normal forty hour work week, and I really see
- 19:28Speaker 1: that kind of a story playing out. In fact, I
- 19:30Speaker 1: can't even think of a client that we're working with
- 19:33Speaker 1: or an organization that I've read about that has implemented AI.
- 19:37Speaker 1: They've trained people, and now there's like there's hard data
- 19:41Speaker 1: to show that employees are working less hours, Like, have
- 19:46Speaker 1: you seen any examples of anyone that's doing this well
- 19:51Speaker 1: as opposed to just saying the right things, because I
- 19:55Speaker 1: feel like there's a bit of a gap there.
- 19:56Speaker 3: Yeah, you know, I think what's coming through and I
- 19:59Speaker 3: have seen this in real life as well, but I
- 20:02Speaker 3: think it's coming through more in the data In a
- 20:04Speaker 3: way that I can stand behind statistically, is that there
- 20:09Speaker 3: is this cluster of use that's really about replacing these
- 20:13Speaker 3: repetitive tasks, these tasks that people don't want to be doing,
- 20:17Speaker 3: and the people who are doing that and choosing to
- 20:21Speaker 3: use the time in ways that are creating enjoyment, positive
- 20:25Speaker 3: experiences at work, more connection. That's what we suffer, this
- 20:29Speaker 3: particular population in our study who had that decrease burnout.
- 20:34Speaker 3: I hear it when people talk about, for example, even
- 20:38Speaker 3: folks who do a lot of interviews, whether it's customer
- 20:42Speaker 3: interviews or media interviews, and they talk about not having
- 20:45Speaker 3: to transcribe anymore and the extent to which they can
- 20:50Speaker 3: really just focus on the conversation and lean into that
- 20:52Speaker 3: and enjoy it and it doesn't have to suck hours
- 20:55Speaker 3: of their time. Those sorts of tasks and the way
- 20:58Speaker 3: people talk about what means to have that time back
- 21:01Speaker 3: and have that depth of interaction. Those are the stories
- 21:04Speaker 3: that I think about and I think are where I'm
- 21:08Speaker 3: seeing the really big lift and so much low hanging
- 21:12Speaker 3: fruit by the way across organizations to find those moments
- 21:16Speaker 3: and those stories and that left and I think sometimes
- 21:21Speaker 3: you know, our job in helping organizations is really identify
- 21:25Speaker 3: those moments where the lift is going to come from
- 21:28Speaker 3: finding that work that can be easily automated and where
- 21:32Speaker 3: the time given back will energize, right versus like be
- 21:37Speaker 3: filled with something that you know will deplete.
- 21:40Speaker 1: Now, you might think of brain fry as just a
- 21:41Speaker 1: personal problem, but that is not the way to think
- 21:45Speaker 1: about it. It is also a leadership one. Coming up,
- 21:48Speaker 1: Gabriella gets into what it actually looks and feels like
- 21:52Speaker 1: when you're in it, how to build self awareness around
- 21:55Speaker 1: your own cognitive limits, just like the way a runner
- 21:58Speaker 1: loans their breath, what leaders can do right now to
- 22:02Speaker 1: check in on whether their team is suffering from it. Plus,
- 22:06Speaker 1: we also get into the ten twenty seventy rule that
- 22:09Speaker 1: most organizations are completely ignoring when it comes to AI adoption.
- 22:17Speaker 1: If you're looking for more tips to improve the way
- 22:20Speaker 1: you work can live. I write a short weekly newsletter
- 22:23Speaker 1: that contains tactics I've discovered that have helped me personally.
- 22:27Speaker 2: You can sign up for that at Amantha dot com.
- 22:30Speaker 2: That's Amantha dot com for.
- 22:36Speaker 1: People listening that feel like I reckon. I am experiencing
- 22:41Speaker 1: brain fry every single day. Like, what are the first
- 22:46Speaker 1: thing or things that you would suggest that they immediately
- 22:49Speaker 1: start to do?
- 22:50Speaker 3: Okay, So the first thing is I got the funniest text, Actually,
- 22:53Speaker 3: my husband got the funniest tax from a friend who
- 22:56Speaker 3: had read our piece on brain fry and was like,
- 23:01Speaker 3: I'm kind of loving my brain fry right now. Like
- 23:04Speaker 3: I'm just like, I know I'm having brain friend kind
- 23:07Speaker 3: of loving He had like all his cloud browsers open
- 23:10Speaker 3: and he was like building a website and he's like, oh,
- 23:12Speaker 3: I'm kind of loving it. So I want to just
- 23:14Speaker 3: recognize that, like these tools can be really engaging and
- 23:17Speaker 3: exciting and often we get to the point of brain
- 23:19Speaker 3: fry because we're enjoying ourselves, right, So it can be
- 23:22Speaker 3: really engaging to see like how much you can build
- 23:24Speaker 3: and how fast, And so the point is not to
- 23:27Speaker 3: deprive ourselves of the excitement and the engagement. And to
- 23:31Speaker 3: my friend, if you're listening, like you go enjoy yourself
- 23:34Speaker 3: and you know, don't deprive yourself of that. The goal though,
- 23:38Speaker 3: is really like this is giving us opportunity to develop
- 23:42Speaker 3: acute self awareness of our own intelligence as we come
- 23:46Speaker 3: to meet this new alien intelligence.
- 23:50Speaker 1: Right.
- 23:50Speaker 3: So I really think it's like this is a deep
- 23:53Speaker 3: self awareness that we can develop of our own brains
- 23:56Speaker 3: that's brought on by trying to understand artificial intelligence and
- 24:02Speaker 3: starting to understand just as when you go out to
- 24:05Speaker 3: go for a run, you start to learn your breath,
- 24:08Speaker 3: you start to learn when am I out of breath?
- 24:10Speaker 3: When do I need to slow down? Those are the
- 24:13Speaker 3: things we have to understand with our own brains as
- 24:16Speaker 3: we're using these tools. We want to operate in an optimum.
- 24:21Speaker 3: We don't want to make mistakes at work, right, And
- 24:24Speaker 3: that was one of the things in our study that's
- 24:26Speaker 3: at stake clearly with this is we need to be
- 24:28Speaker 3: able to work sustainably. We need to have our attention
- 24:31Speaker 3: and our working memory intact for the whole day, for
- 24:34Speaker 3: all the tasks that we have, for the task outside
- 24:37Speaker 3: of work, for our families, right, we want to be
- 24:39Speaker 3: in good shape for all of those things. How do
- 24:41Speaker 3: we pace ourselves? How do we start to feel that
- 24:44Speaker 3: sense of brain fry coming on? Take a break, do
- 24:47Speaker 3: something else, have a digital detox, Start to figure out
- 24:51Speaker 3: how many tools can you reasonably work with at once.
- 24:54Speaker 3: It's probably not more than three. Two, maybe just fine.
- 24:57Speaker 3: One could be just fine too. Find that for yourself
- 25:01Speaker 3: and really, like use an opportunity to learn more about
- 25:05Speaker 3: yourself and your brain and your cognitive capacity. I think
- 25:08Speaker 3: it's a beautiful opportunity for the self discovery of like,
- 25:13Speaker 3: really the most important organ in the body, right, the
- 25:16Speaker 3: organ that determines how we experience Alibliefe itself. And if
- 25:20Speaker 3: this is the technology that's going to get people excited
- 25:23Speaker 3: to learn about it, then amazing all the better.
- 25:26Speaker 1: I do love that idea of just building in more
- 25:29Speaker 1: reflection into your day about how are you feeling and
- 25:33Speaker 1: how is your brain feeling. I remember yesterday I'd had
- 25:36Speaker 1: quite an early start and then I'd been working on
- 25:40Speaker 1: sort of quite cognitively intense work. I've been using AI
- 25:45Speaker 1: throughout the day, and I remember at about like four
- 25:48Speaker 1: o'clock I said to my husband, I think I'm just
- 25:50Speaker 1: I'm done for today, like my brain is done. And
- 25:54Speaker 1: sometimes I have that self awareness and other times I don't.
- 25:57Speaker 1: But I think it's like what you're saying is such
- 25:58Speaker 1: a good reminder to just go. Like at the end
- 26:01Speaker 1: of the day, we are responsible for our own behaviors,
- 26:05Speaker 1: and you know, hopefully you do work in a workplace
- 26:07Speaker 1: where when your brain is feeling a little bit fried,
- 26:10Speaker 1: you can just stop. I would love to know, like,
- 26:13Speaker 1: for leaders, how can they do a formal or an
- 26:17Speaker 1: informal audit of their team to go is my team
- 26:20Speaker 1: suffering from brain fry right now?
- 26:22Speaker 3: I think that, as I mentioned before, being that source
- 26:26Speaker 3: of support, helping the team know and see that you
- 26:28Speaker 3: are there with them, You're learning it with them, You're
- 26:31Speaker 3: there to help them. You know there are no dumb questions.
- 26:35Speaker 3: You will ask those questions if they need to be
- 26:37Speaker 3: asked as well. At the team level, really important that
- 26:41Speaker 3: it not feel competitive, that it not feel like it's
- 26:43Speaker 3: all about who on the team is using the most
- 26:46Speaker 3: versus other people on the team, that it'd be more
- 26:48Speaker 3: about the team pushing one another to help drive the
- 26:51Speaker 3: skills together. As you can do that as a leader,
- 26:55Speaker 3: you can embed tools within your team processes, so it
- 26:59Speaker 3: could be tools that you're using to help digest what's
- 27:03Speaker 3: happened that week, to help drive research that the team
- 27:06Speaker 3: is doing. Doing team trainings together is a great way,
- 27:10Speaker 3: or have team members teach others about how to use
- 27:14Speaker 3: the tools. These are always to help bring the team
- 27:17Speaker 3: together around the tools and to know that there are
- 27:20Speaker 3: new ways that we're all adjusting that are different from
- 27:24Speaker 3: the old forms of work, and keeping an eye out
- 27:27Speaker 3: and open door to listen and to hear about it
- 27:30Speaker 3: and to just learn about what it's feeling like and
- 27:34Speaker 3: looking like as every new generation of tools arrives.
- 27:38Speaker 1: Yeah, I think The issue of metrics is so interesting.
- 27:41Speaker 1: Like I remember in your article it mentioned meta essentially
- 27:45Speaker 1: counting how many tokens staff had used, and the more
- 27:49Speaker 1: tokens the better. Essentially, what are the right metrics that
- 27:54Speaker 1: companies should be looking at when it comes to a
- 27:58Speaker 1: good and impactful use of AI.
- 28:01Speaker 3: We're in a stage now where the emphasis is on
- 28:05Speaker 3: really getting the most out of the tools, not from
- 28:09Speaker 3: a quantity of activity perspective, but really from an amount
- 28:13Speaker 3: of value that's being generated, right, and so how do
- 28:16Speaker 3: we think about the quality of the outputs and the
- 28:19Speaker 3: value of the outputs, and the value of the engagements
- 28:22Speaker 3: and the experiments and you know, we like to talk
- 28:25Speaker 3: about that in our world of course, as learning as
- 28:29Speaker 3: the rate of experimentation. Those are all great ways to
- 28:33Speaker 3: try to capture how this is being metabolized effectively through
- 28:38Speaker 3: the organization. Any kind of change in an organization has
- 28:42Speaker 3: a toll. So before we want to roll something out
- 28:46Speaker 3: and have lots of people using it, we want to
- 28:49Speaker 3: be really thoughtful and strategic about how and why, and
- 28:53Speaker 3: so being really careful about how we design around all
- 28:57Speaker 3: of this is the other piece of what we see
- 29:00Speaker 3: happening right now in the phase that we're in now,
- 29:02Speaker 3: and the more we can carefully design how work should
- 29:05Speaker 3: be flowing, where it should be locating, how the tools
- 29:09Speaker 3: should be used, how many tools right on, what spans
- 29:13Speaker 3: of control those types of questions, the more we can
- 29:16Speaker 3: make sure that the changes we're going to roll out
- 29:19Speaker 3: will be to the right people and the right ways
- 29:21Speaker 3: at the right time and kind of minimize that change
- 29:25Speaker 3: burden that over otherwise will create more inertia than forward momentum.
- 29:31Speaker 1: Love that now. Something I'm experimenting with on how I
- 29:34Speaker 1: work is finishing with what feels cliched, but some rapid
- 29:38Speaker 1: fire questions that I have for you, So if you're
- 29:42Speaker 1: out for it, my first question is which industries will
- 29:46Speaker 1: feel this in terms of brain fry and the associated
- 29:49Speaker 1: problems the hardest. Do you think in the next twelve months?
- 29:52Speaker 3: Okay, So in our data set, the brain fry risk
- 29:57Speaker 3: was greatest for the most operational and the most technical,
- 30:01Speaker 3: and so I think that it would be the roles
- 30:04Speaker 3: that tend toward those things the most. But I think
- 30:08Speaker 3: that it will be more about whether the organization is
- 30:12Speaker 3: making the right design decisions around how to be using
- 30:16Speaker 3: the technology, regardless of industry. These are avoidable challenges, if
- 30:22Speaker 3: we make smart decisions around the design, if we learn quickly,
- 30:26Speaker 3: if we help empower employees to take pauses, have leaders
- 30:30Speaker 3: who are there to answer questions. I think it's actually
- 30:34Speaker 3: more about creating cultures where those things are possible, making
- 30:38Speaker 3: decisions that are driven from the right outcomes, driven from
- 30:42Speaker 3: a desire for sustainable productivity value creation.
- 30:46Speaker 1: What is a non negotiable that you have in your
- 30:48Speaker 1: life to protect your own cognitive capacity?
- 30:51Speaker 3: I mentioned the digital detox. Another one is time with
- 30:55Speaker 3: my family, So I tried to have dinner with my
- 30:59Speaker 3: family and do a bad time. I'm every single night
- 31:01Speaker 3: with my kids, very grounding, centering about the things that
- 31:05Speaker 3: matter most to me. And you know, I think one
- 31:10Speaker 3: of the main drivers of resilience in my research across
- 31:14Speaker 3: large populations is self compassion and as you well know,
- 31:20Speaker 3: part of what that means is keeping it in perspective.
- 31:22Speaker 3: And there's nothing like you know, kids to help me
- 31:27Speaker 3: keep things in perspective.
- 31:29Speaker 1: Yeah, I love that very much. For light, what does
- 31:31Speaker 1: a healthy relationship with AI look like?
- 31:35Speaker 3: In twenty twenty three, my lab that I was leading
- 31:38Speaker 3: at the time did a study to predict who would
- 31:43Speaker 3: be most inclined to use AI tools, and it came
- 31:46Speaker 3: down to a high degree of optimism and a high
- 31:49Speaker 3: degree of agency. And that's what I would point to
- 31:52Speaker 3: an answer to the question is that the more we
- 31:55Speaker 3: can be optimistic about how we can use these tools
- 32:00Speaker 3: to advance our goals, to advance our purpose, to advance
- 32:03Speaker 3: the things that we want to accomplish, and the more
- 32:06Speaker 3: we can lean into our sense of self efficacy and
- 32:08Speaker 3: agency that we are in the driver's seat, the more
- 32:11Speaker 3: that we can use them to our advantage and really
- 32:14Speaker 3: get the most out of them as individuals, as leaders,
- 32:17Speaker 3: and as organizations.
- 32:18Speaker 1: Okay, what's the single biggest mistake that organizations are making
- 32:22Speaker 1: with AI adoption right now?
- 32:24Speaker 3: We have seen over and over in our research that
- 32:29Speaker 3: the formula of getting adoption right is ten twenty seventy
- 32:34Speaker 3: ten percent. A way of to get right is the algorithms,
- 32:37Speaker 3: twenty percent is the data and the tools, and seventy
- 32:40Speaker 3: percent is the processes in the people side of things.
- 32:43Speaker 3: What organizations often miss is that seventy percent and the
- 32:46Speaker 3: focus on people. So really focusing on the people side
- 32:51Speaker 3: of getting people to find that high agency, high optimism
- 32:56Speaker 3: mindset to engage with the tools to use it in
- 33:00Speaker 3: ways that authentically advance their goals, their purpose, that give
- 33:05Speaker 3: them that sense of lift that we talked about, to
- 33:08Speaker 3: have more time for connection, for belonging, for purpose, for creativity.
- 33:13Speaker 3: Those are the things where we really see this like
- 33:16Speaker 3: unfair advantage opening up and huge opportunity.
- 33:21Speaker 1: And finally, if you had one minute with the CEO
- 33:24Speaker 1: of a large company, what would you tell them about
- 33:26Speaker 1: brain fry?
- 33:27Speaker 3: I would say AI is completely changing the psychology and
- 33:33Speaker 3: behavior of work. Brain fry is a proof point of
- 33:37Speaker 3: that we need to learn quickly and use all of
- 33:41Speaker 3: this information to redesign work for the benefit of our employees,
- 33:46Speaker 3: of our leaders, and of our bottom line.
- 33:48Speaker 1: Love that, Gabriella. Thank you so much for coming back
- 33:52Speaker 1: on how I work. I love dat chat the first time,
- 33:55Speaker 1: and it's just been so great to dive into your
- 33:58Speaker 1: research about brain fry. I've been thinking about it so
- 34:01Speaker 1: much since I read that piece in HBr, So thank
- 34:04Speaker 1: you for continuing to just put such great work into
- 34:06Speaker 1: the world.
- 34:07Speaker 3: Thank you so much for having me back of such
- 34:09Speaker 3: a pleasure to be with young.
- 34:10Speaker 1: So when you think about the most sophisticated AI users,
- 34:14Speaker 1: they're not actually the ones with the most tabs open.
- 34:18Speaker 1: The people who figured out AI are the ones who
- 34:21Speaker 1: figured out themselves, their own cognitive capacity, their own pace,
- 34:25Speaker 1: their own version of two tools versus five. So it's
- 34:30Speaker 1: less about the technology and more about self knowledge. Now,
- 34:34Speaker 1: if you are someone who ends the day with a
- 34:37Speaker 1: fried brain and no real sense of why, try treating
- 34:41Speaker 1: it like Gabriella suggests, like learning your breath on a run,
- 34:46Speaker 1: notice when it starts, and take the break before you
- 34:49Speaker 1: actually need it. And if you enjoyed this chat with Gabriella,
- 34:54Speaker 1: I reckon you'd enjoy the first time I had her
- 34:56Speaker 1: on How I Work, where we talk all about thriving
- 35:00Speaker 1: in times of uncertainty. If you like today's show, make
- 35:04Speaker 1: sure you hit follow on your podcast app to be
- 35:07Speaker 1: alerted when new episodes drop. How I Work was recorded
- 35:11Speaker 2: On the traditional land of the Warrangery people, part of
- 35:13Speaker 2: the Kohen Nation.