AI Is Making Work Sadder: And What Happens When Everyone Shares the Same Brain?

What if the most disruptive thing about artificial intelligence isn’t that it takes your job, but that it makes your company indistinguishable from every competitor — and drains the humanity out of the work that’s left?

Aaron Horwath has an unusual vantage point on that question. As director of AI operations at a creative technology company, his job is to put these tools to work. But in two essays for the longform magazine Noema, published by the Berggruen Institute, he’s become one of the sharpest voices asking how much the corporate rush to “use AI for everything” is costing us as human beings.

In this conversation, Horwath explains why even senior professionals with decades of experience are feeling a strange career dread. More than fear of unemployment, it involves the loss of something human: the ownership, the craft, the pride in work that was recognizably their own before AI wedged itself into the workplace. 

He traces the sameness spreading across corporate output to a deeper problem: When nearly every company runs its thinking through the same two or three large language models, those normalization machines strip away the quirks and perspectives of the very people companies pay to hire, and everyone ends up with the same ideas, the same solutions, the same voice.

But Horwath moves past the hand-wringing to how businesses will actually have to respond — by putting humans first. He predicts a “barbell” economy, split between high-volume machine output and premium, deliberately human work. He points to companies that now bar new hires from using AI for a year so they learn to think, foresees coming fights over who gets to be a tastemaker and the threat to apprenticeship, and explains why the smartest firms are already carving out AI-free zones.

His closing insight reframes the whole moment: For decades we bent technology toward making humans more productive. The real task now is making the workplace more human. Listen to hear why that may be the defining business challenge of the AI age.

Apple PodcastsGoogle Podcasts RSS

Full Text Transcript:

Jeff Schechtman: Welcome to the WhoWhatWhy podcast. I’m your host, Jeff Schechtman. For most of the history of technology, the bargain has been pretty straightforward. A new machine comes along and it does something faster, cheaper, or better than we can. Some jobs disappear, others are created, productivity goes up, and eventually we adjust. Artificial intelligence may be offering us a rather different bargain, because what if the real disruption is not simply that AI can do our work, but that it begins to change what work actually is? What if the lawyer, the marketer, the programmer, the consultant, the writer, all those people we once hired for judgment, experience, taste, and imagination slowly become managers of machines that do most of the thinking for them? And what happens when everybody is using essentially the same machines?

We have spent years worrying that AI might make human beings obsolete, but there may be another possibility hiding in plain sight: that long before it replaces us, it makes much of what we produce look and sound exactly alike. The same language, the same ideas, the same solutions, the same corporate voice. An enormous increase in output accompanied by a quiet collapse in originality.

And then there’s an even more uncomfortable question. If the machines remove the drudgery, the meetings, the arguments, the false starts, the hours spent wrestling with an idea, all the inefficiency we’ve spent decades trying to eliminate, what if they also removed the very things that make work occasionally satisfying? Maybe friction was doing more for us than we realized.

My guest, Aaron Horwath, has been thinking about all of this from an unusually interesting perch. He is the director of AI operations at a creative technology company, which means his job is, quite literally, to help bring this technology to the workplace. But in two provocative recent essays for Noema, he has also been asking whether the great corporate rush towards efficiency may be quietly producing something else: intellectual sameness, diminished skills, hollowed-out careers, and an entire generation of knowledge workers beginning to wonder what exactly all this work was for in the first place.

So this is not really a conversation about whether AI is good or bad, or whether the robots are coming for your jobs. It’s about something older and perhaps more important: what human beings need from work, what organizations actually get from human beings, and whether in our eagerness to strengthen every bend in the river, we may discover that the bends were where some of the most valuable things were happening. It is my pleasure to welcome Aaron Horwath here to the WhoWhatWhy podcast. Aaron, thanks so much for joining us.

Aaron Horwath: Oh, thanks for having me. I’m excited.

Jeff: Well, it is a delight to have you here. I want to begin before we talk about technology and what it all means about this idea of sadness that you talk about in the first of your pieces, and you begin the piece with a young executive sitting on a train, talking about finance and EBITDA and all sorts of financial-related things, and suddenly he takes out knitting needles and shows somebody a hat that he’s knitting, and suddenly he lights up — he’s excited about it. It’s a whole different personality, almost. Talk about that, where that notion came from ,talk about that moment.

Aaron: I think it was a super nice metaphor, or at least a symbol of something that I was certainly starting to feel, and I was noticing people starting to talk about around me as well. It was two things simultaneously. One was even people more senior in their careers were starting to, it seemed, get a little bit uneasy with maybe what the future of their careers looked like and what AI was going to mean. There was also, as I say in the piece, just kind of a sadness for some reason, and I wanted to get to the bottom of that. And I think also just the kind of a self-exploration around what I was feeling and what other people were feeling, because the other thing that people kept coming up talking about was this desire to get back to these old offline analog tasks, or crafts. So people getting into knitting and writing again, or painting or some sort of craft and feeling a desire at the end of their workday to do something that felt, quote unquote, real. That seemed to be common amongst a lot of people that I was speaking with. So I wanted to get to, through the piece, get to the bottom a little bit of what I was feeling and what my peers were feeling as well, and what was driving that impulse.

Jeff: We see this huge explosion in live events today, particularly sports and concerts and the like. Is that somehow related to this, the idea of people being more connected with something in person?

Aaron: I think — particularly when it comes to the knowledge work part — one of the things I identify in the piece as potentially driving this feeling is the abstraction, or the sense of abstraction that arises when you start to really leverage some of these AI tools, particularly in companies that are pushing for employees to use AI in every instance possible. So kind of an AI-for-everything strategy. After the piece came out, I heard a lot from developers who felt like they were suddenly spending a day that they used to spend writing and crafting code, all of a sudden just hitting enter in Claude and attending meetings, and then also other people who felt like maybe the work that they did in knowledge work was not super altruistic. They weren’t nurses and they weren’t school teachers, but they did feel some sense of pride over the craft that they use to make a living. And AI had come in and put itself between them and that craft in a way that felt a little bit too far in some sense, like it was a step too removed from the actual execution of the work and feeling any ownership over it at all.

As you mentioned, the opening, the idea that what’s produced when you put AI between the human and craft can often feel very similar. And so there’s also this sense from employees maybe that I’m not even hired anymore because of my unique talents. Basically, anybody can prompt Claude or prompt ChatGPT and get an output. So what is actually my value and my unique role within an organization?

Jeff: And this relates to this idea that you touch on that Derek Thompson had talked about, this idea of workism, that so much of a job today is not just a job anymore, but it is connected so much with our status, with our identity, with our self-expression, etc.

Aaron: It’s a good metaphor, religion, because I think there’s work, and then there’s all of the things that many people enjoy about the work, which is the community, collaborating with people — hopefully some of them even that you like — solving problems, maybe getting to work with people from different cultures, all of the things that make work meaningful.

And we used to, pre-AI, you had this work process where you would have an idea, and then you would need to bring that to your boss, and maybe they would give you some feedback on that idea. And then you would take it around to colleagues and get their input. And the end result was, one, a process that was not super efficient, but it allowed the opportunity for collaboration. But it also meant that the product of that work was unique to you and your organization, and all the micro-interactions in the collaboration that happen along the way also benefits the business in many ways — but also adds some sense of meaning to the work itself.

I can now basically sit down as one person and have an idea and push it all the way to done. And I think there was some excitement early on about the fact that that would allow organizations to move faster. But I think for the employee experience and for businesses, there are some risks that are worth taking into account when you cut all of that out of the process, kind of like that meandering, less efficient but very innovative process.

Jeff: To what extent do you think that this is really about AI, or did AI simply arrive and come to us at this moment when a generation of workers, knowledge workers in particular, really wanted to admit that something is fundamentally wrong?

Aaron: I think that’s probably fair to some extent. But the things that I thought were a little bit different were, again, one, it wasn’t just people coming out of college who were nervous about their career and getting into the job market. It was also people who had 10, 15, 20 years and were fairly senior who were expressing real angst and kind of career existential worry about their work and their future career and what work they would be doing. So to me, that felt quite different.

And then the second is that AI, an AI-for-everything strategy where you’re trying to use it as much as possible in every single instance, does create almost a manufacturing-like work environment, which is not really what knowledge workers have signed up for, is to push out volume repeatedly over and over again. Most of them have come to their careers through interest in, like I said, problem-solving, collaborating with people, innovating, coming up with new ideas, being creative. If a lot of that is removed also, then it gets even harder for them, I think, to justify their career in something like knowledge work. And so I think it does uniquely beg an existential professional question that people were and are currently trying to answer for themselves.

Jeff: I guess the flip side of that is that if AI disappeared tomorrow, if ChatGPT and Claude disappeared, would any of these people be any happier at work again?

Aaron: I mean, I don’t want to make too broad of a statement. And I think also in some ways they would be less happy. I mean, there are really good use cases for AI. There is a lot of work that knowledge workers do that is sort of quote unquote meta work. So filling out your HubSpot or your CRM or finding that clause in that contract that you need to find. There’s lots of work that people do in preparation for actually executing the role that they were hired to do.

So internally, at Creative Force, our focus has been from the very beginning to focus AI on those areas, the work that people aren’t motivated to do, but that also is critical, but isn’t the most motivating work. And so our goal has been, could we use AI to take care of some of that drudgery so that our people can focus on the work that they find most motivating, but also that drives value for the business?

Like a salesperson doesn’t go into sales because they’re super excited to fill out their CRM or HubSpot. They are excited about working with prospects and going to events and working deals and helping people purchase whatever the product is that they’re selling. That’s the stuff that makes them most motivated. As a business leader, it’s also the thing that business leaders and executives want their people spending their time on. And so that’s how we’ve framed and focused our efforts. So in some ways, if you took away those tools, it would also mean that a lot of that drudgery comes back to the employee experience.

Jeff: How much of the meaning of work, I wonder, has to do with compensation, and that sometimes the work feels more meaningful to people that are being paid a lot of money to do it.

Aaron: Oh, that’s an interesting point. That’s not where I thought you were going to go. I thought you were going to say that there’s something about the inefficiency, but like the work that goes into something that gives it meaning. So, for example, there might be a project that two years ago would have taken us three months and 20 people and 15 meetings for us to deliver, and today two people could deliver it in a few hours. That’s a real example that we have in our org.

But the way that we did it two years ago, one, I think just the fact that it took so much effort to do it made it meaningful in itself. But then also all of those meetings, the learnings that we got from the process and the conversations that we had, our organization also benefited from all of that. And today, if we weren’t conscious in how we were using AI, we would lose all of those benefits because I could just sit here and basically create it right now and push it live, and there you go. But I think organizations will need to think a little bit harder about, if we introduce AI here, what are we gaining, which we’ve been really focused on, but also what are we losing by doing so?

Jeff: This is this messy middle that you talk about, that inefficiency is fine to want to take out of the system — too many meetings, too many people, too many arguments. But in fact, by losing some of the friction of that inefficiency, we lose an awful lot.

Aaron: Yeah. And I think in some cases, the friction is — it’s good to get rid of, but I think it’s not as simple as, let’s get rid of all friction so we can move as fast as possible.

So as a very practical example, we were just talking yesterday in the org about using Claude and AI to set up demo accounts for prospects who are interested in our software. We can have AI probably do a pretty good job at setting up a demo account so that our people don’t have to spend time doing that. That’s a very repeated task. But there’s two interesting things there. And one of them is that by setting up a demo account, they learn about the prospects and their needs and what they’re currently doing and what their wishes are. So there is a learning there.

And then two, if we had Claude set up a demo account for them overnight so that when they woke up, it was ready, they would still need to review it. And the question is, would they spend as much time looking through what AI had done to learn it and make sure that everything was working as if they had started from scratch? Still up for debate, but that’s a much more complicated — or it’s a much more nuanced approach to AI than what we’ve been doing. Not we, but like generally, the conversation in tech has been use AI in every instance possible. That has been the approach of a lot of organizations.

And I think that it’s going to need to be, and it will be, a more nuanced conversation in the very near future as people start to think about, well, we introduce AI there and we automated it and we made it faster, but did we lose something along the way by doing so?

Jeff: Which really gets to this notion that, arguably, is at the heart of all of this, both the happiness aspect and the work-related aspect, that in fact, work is not just a production system. It’s not just an assembly line, but it really is a larger social system.

Aaron: And I think that’s the business model. That knowledge work is predicated on that we’re going to hire and pay pretty well a bunch of unique people with unique perspectives and opinions and quirks and experience. We’re going to put them in a room, and the hope is that they’re going to come up with solutions that are unique enough and innovative enough that our customers are going to pay for. That is basically the business model.

And so if you decide that you’re going to turn it more into a manufacturing environment, that’s not a great environment for innovation. You don’t want the guy on the assembly line building cars failing fast on how he connects the brakes to the car. You just want them to do the same process over and over again.

You add to that that most companies — I mean, there’s only two or three of these tools that companies are using at scale. And so then you’re also taking all those unique quirks and perspectives and experiences, shoving them through the same LLMs, basically normalization machines, and then getting the outputs, the same outputs, the same solutions as your customers have access to, and also that your competitors have access to. And that kind of erodes the knowledge work business model from the inside out, I think.

And so I think people are going to need to be very careful in making sure that they’re protecting their competitive differentiation because everyone has access to these tools. So then the question is, in an AI world, how do you leverage AI where you’re getting some of the benefits, particularly for the employee experience, and so they can spend time on value-generating activity in the org and that they find motivating, but also that protects your competitive advantage and kind of your magic, your magic sauce that makes your company and your products and your solutions unique.

Jeff: You talk about it in one of the pieces with respect to sports analogies, that sometimes we just have to let people be bad at things, that an umpire can make a bad call, or that being bad at something sometimes has a broader positive result.

Aaron: There’s just something a bit deflating at times where you have a — and this is not just content-focused; it’s also business strategy and problems within an org. You put a problem into Claude and/or ChatGPT and you get an answer out. And whereas previously we would maybe go through some iterations and we would get some learnings and we would work with a group of people on it, there is something that, again, like we talked about before, makes it feel a little bit cheap to just put in a question and get an answer out, copy-paste it into a document, and then boom, there’s your strategy for your team or whatever it is.

I think the fun of work is that process of trying to figure things out, particularly with other humans. And I would — I’m wary about removing that completely from the experience of work for people.

Jeff: Do you think that that’s a real danger as you see the business world evolving, that we have a world where people really are only interacting with their screens, with their AI?

Aaron: I think that there will be — the reason that I wrote the two pieces, the first one, Why Is Everyone in Tech So Sad?, was very much like an emotional zeitgeist piece for employees. But the second piece, Now Entering the Era of Blegh, was a piece making the business argument that there is going to be incentive for businesses to protect their competitive advantage by making sure that they keep some of that messy middle that protects their differentiation in the market.

So if companies who start to think about this really seriously are just going to be much more targeted in how they leverage AI, I think they’re going to — and the work that they’re going to need to do, in all honesty, is going to be on a task-by-task basis. Because every time that you introduce AI to a workflow or a task or a process, you’re making a strategic decision about that work, or you’re making a strategic decision about the role that AI is going to play in that work.

So a common example that I use, just because it illustrates it very clearly, is: A company, any company, has a newsletter, and for the last couple of months, they’ve probably been writing most of it with Claude or some AI tool. They’ve had their agent go out to the internet, get industry news, bring that together, write it in the format that they like. That’s a high-volume, low-differentiated approach to a newsletter. They’re probably hoping that they could pump out more newsletters more often. And it’s kind of an outbound play, but it’s not super differentiated. I think organizations are going to start to look at things like that and say, okay, we could do the high-volume, low-differentiation approach to this, but we could also hire a copywriter with really strong taste, with a voice that’s really recognizable, who’s really uniquely good at conveying this type of information we want to convey to our industry, and we’re paying them to actually not use AI at all. We want the most differentiated output possible in this instance.

The newsletter is just an example, but I think that there could be in the future a barbell curve that forms where you have value generated on one of two barbells: on one side, that high-volume, low-differentiated work, and then on the other end, highly differentiated, low-volume work. And you’re going to see maybe a possible elimination of the information in the middle — or not the information, but that kind of middle work will be eroded out.

Jeff: In many ways, it’s reflective of so many other industries. I mean, I think of the fashion industry, for example, which is a low-end business on one end and a very high-couture business on the other end, or even cars, where you have a huge difference between the average vehicle and the luxury market. We see that happening in so many other areas.

Aaron: Yeah. And I think the question becomes for knowledge workers, is for a little while right now, I can set myself apart by using AI in my role. But long-term, that’s going to be a bit like saying I’m really good at spreadsheets. I mean, there are some roles that are specifically — maybe you are really good at spreadsheets — but it’s just going to be a tool that people use in their work.

And so they will have to make a decision about, or be more active about, how do I get myself into that far-right barbell curve where I have a recognizable perspective that’s highly differentiated, that basically shows to an organization, I can help you in the places where you need it to stay outside of the AI-generated pipeline, both in terms of ideas and content, and make sure that we’re not falling into kind of a non-differentiated, just AI-generated content, ideas, solutions, etc.

That’s, I think, going to be increasingly maybe an emphasis that knowledge workers are going to have to put on themselves around figuring out, how do I establish myself as one of those people? Because we’re seeing it already at the big AI firms, is that they are hiring these people with really high salaries to be in these tastemaker roles. So they obviously also recognize that this is going to be a skill and a knowledge area that has a lot of value in the future for them.

Jeff: It also goes to the whole business of recruiting at companies. Companies spend a fortune on HR and recruiting, trying to get really great people and a diverse mix of people and different experiences, different instincts. It would be so useless if all of that goes towards putting the same output out through AI.

Aaron: Yeah. And that’s what triggered the second piece for me, was just that I thought that was quite funny, that that had been the way that we had gone. And I don’t think that will be long-term, but it just seems odd for companies who are dependent on unique perspectives and experiences to be so excited in the short term to shove all of that into the big normalization machines of LLMs. But I think that will be — I would assume that will be not a long-term situation.

Jeff: It relates to this idea you call airbrain in the piece, really, that we start to homogenize virtually everything.

Aaron: Yeah. That came from two places. One was obviously everybody now knows that em dash’s life is on life support right now because they are seen as a primary sign that something has been written by AI. But also the paper or the article that I cite in the piece by Kyle Chayka — his is called Welcome to Airspace — and he wrote it in 2016 and described how algorithms and social platforms like Instagram and Foursquare had created an aesthetic homogenization across the globe, particularly in places like cafes and bars and Airbnbs. So if I was standing in a cafe in Tokyo, I wouldn’t know if I was in Tokyo or if I was in Berlin or if I was in Sacramento. They all kind of had this homogenized aesthetic that had been driven by the social platforms. And when we started to see, all of a sudden, all of these em dashes pop up in everybody’s content online, I saw a real connection there in terms of like, hmm, what if we start to see a similar homogenization not in physical space, but in intellectual space, and people starting to come out with not very differentiated, not only content, but also solutions and thought process.

Jeff: It really makes originality so much more powerful and so much more valuable because ordinary competence just becomes a commodity with AI.

Aaron: I don’t know if you feel the same way or how often you’re writing, but for me, I find myself pushing my tone and my style even harder in a particular direction to make sure that I’m staying outside of that lane. One of the really nice things people said, particularly about the first piece, was that they really appreciated the writing, which obviously means a lot. But also I — and thank goodness I have an editor, or I had an editor at Noema, because it was even crazier in the original version — but I really was trying to push the tone and the style as much as possible, just to make sure that it stood out amongst what is — not at Noema, but elsewhere, just on the internet — lots of AI content.

Jeff: Do we get to a point where content and things start to get labeled “made by human” or, you know, the same way things are handmade or organic or small-batch now?

Aaron: I will not predict the future, but I can certainly see a premium being placed on that. I think also places like Substack, obviously people can still use AI on there, but this idea that you have some sense of the person who is producing the content already seems to have some level of value, at least in terms of the writing side of things.

Jeff: I want to come back to this idea that one of the things that gets eliminated in all of this is — and you talk a little about this — the idea of juniors, young people coming up. If they get eliminated, how do we get to higher thinking in many of these business environments?

Aaron: I think it’s particularly a concern for the consulting firms because so many of those bottom rungs of the organization are really preparatory roles where they’re doing a lot of work that Claude or AI is pretty decent at: doing research, putting together documentation, putting together slide decks. But obviously, those roles have always been with the assumption that you’re going to learn throughout that process so that you are ready for when the day comes that you’re going to get promoted into a more senior role. I don’t think it’s only in those firms. I think SaaS has a similar issue, which is that there are only so many senior people. So I could imagine also, and I’ve seen at least one company doing it already, which is saying, for your first year at the organization, you’re not allowed to use AI. You’re going to learn our tool by manually using it and exploring it yourself. You’re going to learn about our industry by talking to people and doing the manual work of getting to know the industry, etc.

And I could really see that being something that people start to introduce, because otherwise, I think you do end up with a bit of a hollow expertise where people maybe make it to a certain rung, but once you throw them into the meeting room where they need to have that face-to-face conversation with someone, and you can’t just pull out AI to answer a question, they may not have the skills that it may have seemed that they had sitting behind their desk.

Jeff: It eliminates, potentially, this whole idea of what we know today as apprenticeship, whether it’s in a law firm or in something else.

Aaron: Yeah. I think it is one of the places where people who have been very excited, particularly business leaders who have been excited about AI, should keep an ear open to people like their people teams or their HR teams, because I think those are — that’s the team that our AI initiative lived in. And I think that was super important. I think a lot of organizations either have a very top-down approach where their executives are really driving AI adoption through some pretty superficial metrics, or they’re handing it to their IT teams — and their IT teams should be involved and their executives should be involved — but it is a team, kind of like a people team or a COO or others who are also thinking about some of the more work-philosophy questions around: If we introduce this technology, what does this mean for our talent pipeline? What does this mean for the skills that we need? How does this change our roles? You know, in the past, you might have junior salespeople whose job it is to send a lot of emails out. But if AI can do a lot of that work, it doesn’t mean that you get rid of your junior salespeople, but it might mean that they’re now kind of the people who go out more often to industry events and have more of those human-to-human interactions. It just changes the dynamic of the role, not the need to eliminate the role itself.

But your IT team is maybe not super well positioned to make those judgment calls. They could certainly have an input, but that is very much a people team or an HR team role. And I think that’s where these projects should probably most often sit.

Jeff: It also raises the question of who gets to be the tastemakers in this kind of environment.

Aaron: That one remains to be seen. And what people are going to be valuing, is it that — I mean, one thing we know, EY recently announced that they were allocating $100 million to celebrate and reward uniquely human and innovative work within their firm. I don’t think that every organization will go for, like, a monetary reward, but I think that that’s super interesting that there is already this focus on who can come up with solutions that Claude isn’t typically going to output, who can come up with the most unique-sounding content or the unique solutions or the thing that stands out? I think there will be a real premium on people who can demonstrate that they’re able to do that regularly.

Jeff: And do the models, I wonder, get good at taste at some point?

Aaron: I hope not. No, I mean, maybe they will. I think there’ll be, certainly, versions of it. You can imagine at least organizations having their own tuned — and some are already doing it, of course — but having their own tuned LLMs that are trained on their historical data and get a little bit closer maybe to content and solutions that fit the org. But again, I think you’re still in a situation where you need people to — all of that is based on historical data — and you need the people in your org who are thinking a little bit forward, forward-thinking about new solutions that are outside of that AI pipeline.

Jeff: The other part of this, and we’ve been talking about it pretty specifically with respect to enterprise, but a lot of this is applicable to human relations in general as well, and how that potentially changes from all of this.

Aaron: Well, certainly. I mean, in the workplace, it is another place where I think organizations are going to need to put some effort, because it is very easy to sit at your desk as a new employee and get a lot of the information that I need in order to do my job from agents. Certainly within Creative Force, we have hundreds of agents that can give you a lot of information about our customers and our industry and our tool and all the rest of it. Whereas during your first three to six months, you’d be spending a lot of time with people just asking a lot of questions.

Now, are the answers that the AI is giving you as nuanced and specific and accurate as the ones that senior people would give you? Probably not. But it’s certainly a lot easier to ask the agent. And so we’re also going out of our — we go out of our way for our new employees to make sure that they’re getting around the org, meeting people, not face-to-face, but at least virtually face-to-face, and having those conversations, even though they could, in theory, only ask agents.

But it is something that I think organizations are going to need to invest in, is making sure people aren’t locking themselves away alone, just figuring it out and still learning from people who are more experienced or have been at the company for a long time or have that unique knowledge that they need access to.

Jeff: How much did remote work and the rise of remote work, particularly during the pandemic, change the environment in a way that makes all of this seemingly more possible today?

Aaron: I don’t know. Maybe there’s a connection with AI, but it seems that there was more just a switching capability at some point. End of 2024, it was clear that the tools could actually start to really do some work. And then it took a little bit longer to figure out — and I think many organizations are still trying to figure out, whether you’re remote or not — what is the best way to integrate this technology, and what is the best way to get people to use it?

And that’s where most people stop. And I’m sort of pushing on. There is one more step, like one more step to take, which is not just what can the technology do, but what should it be doing and how should we be using it? And that’s, to me, a much more interesting question and much more nuanced. And it’s something that every organization at some point will have to navigate, I think.

Jeff: And if we don’t navigate it, then it becomes, one, either very problematic for what the results are relative to what we’ve been talking about, or the need to unwind it later. You use the metaphor of the river, the Kissimmee River, and that is a perfect metaphor for what we’re talking about right now.

Aaron: Yeah. So just for context, the Kissimmee River in the 1960s in Florida kept flooding. And the reason was because there was like 200 miles of winding bends and kind of these outflow portions of the river. So the Army Corps of Engineers decided that they wanted to straighten it to increase the flow of water, which worked. But by straightening all the bends, you also saw an increase in the death of good fish, increases of garbage fish, bad aeration, silt going directly into the main waterways. There was a lot of downsides. And so they ended up building back in some of the bends for like 100 miles of the river to get back some of those benefits. The reason I use the example in the piece is because I think there will be a time for every organization where they will start to do the exact same thing. They will start to build back in the bends of innovation, where they say, actually, we said we wanted to use AI for everything, but let’s take a more targeted approach to how we’re integrating this within our teams and the work that we are doing.

And that’s going to take a lot of work. Like I said, I think that’s a task-by-task project, but it is the work that’s going to need to get done in order to make sure that you are getting those benefits that AI can provide your employees and also your business, but also making sure that you’re not succumbing to some of the downsides that it can present as well.

Jeff: Are companies going to need to have AI-free zones, essentially?

Aaron: Super interesting that you mentioned that. That is also one of the things that we are talking about as well, is if there will need to be more formalized projects or instances where we don’t use AI at all, particularly in the ideation stage. Definitely there will be tasks that teams will decide: This is not something that we’re going to use AI for, because we think that we want to rely on the innovation and the uniqueness of our employees to come up with a unique solution. Maybe they’ll use AI a little bit, but generally, like, this is an AI-free task that we think offers us a competitive differentiation in the market, and we don’t want AI touching it.

There will be other tasks where organizations will say, we want to use as much AI here as possible. It doesn’t impact differentiation, and it’s a great efficiency gain for us. And so that would be kind of like on the work-mapping side.

And then definitely on the cultural side, like I said, do we say for a year some of our new employees can’t use AI? Is there certain projects that we want to just say, hey, let’s not use AI for this. Let’s go old-school on it to make sure we’re getting a differentiated solution here. I think those conversations are the type of conversations I would suggest companies start having.

Jeff: It’s so interesting that we have spent so many years through the evolution of technology trying to figure out ways in which technology can make human beings more productive. And now we’re at a point where we’re trying to make human beings more human in the workplace.

Aaron: I think that’s the fundamental — that is the fundamental driving motivation that we had also, was that we — the goal was never, and our executives never said, that we want to remove headcount; they didn’t even really talk about efficiency. Our story from the very beginning was we want to make sure that our employees are spending the time on the work that they find most motivating, and we believe that the work that they find most motivating is also the work that drives the most value for the business.

If we can use AI to give them back some of that time so they can spend more time on the work that makes them excited, that’s going to make their performance better. And it’s going to also mean driving more value for the business. There is, I hope, a very — you’re going to have to walk a fine line, but I think there is a way here to leverage the technology where everybody wins. And that’s what we’ve been promoting within Creative Force. And I hope more orgs start to look at it in a similar way.

Jeff: How much harder is it to do that, given that the technology is changing and improving and getting better, more aggressive with each day?

Aaron: I think the technology can get as good as it wants. It’s up to us to make the decision that, as an — you know, if you were a business leader or whatever — that there is a way that we want to use this technology, and there’s a way that we’re not going to use it. And this is our general stance or philosophy around how it should be used. 

Then there’s definitely some work from that people team and your AI enablement team that you put together with stakeholders from around the org to go out to people leaders and individual contributors and help them think through their own work to say, where do you see AI helping you, making your experience work better, giving you more time to work on the things that you care about, but also make sure that we’re protecting the thing that makes us unique in the markets? That part will take the most work. It is always the people work that takes the most effort. We can build an agent in two minutes. That part is very easy, and you can build hundreds of them. But it is the people side that takes the most effort, but it’s definitely worth the investment.

Jeff: And do you see that becoming a part of HR going forward, or companies hiring individuals to specifically navigate this landscape we’ve been discussing?

Aaron: I think it will depend on if people think it’s a full-time role or not. Obviously, that is my full-time role, and it has been for two years or so now. I know that I’ve talked to people in IT teams who are also not interested in owning that part of it, particularly the change-management part and the work-design parts. You know, that’s not an area that they’re interested in. So I do see, and I’m increasingly seeing HR and people-team roles mentioning this type of work around AI enablement and work design and process design and things. So I think it will ultimately live and ultimately should live in those teams.

Jeff: Finally, how seriously are people in the corporate world and enterprise taking all of this? Is this just a discussion right now, or is this, as it is in your company, something that really is taking up a significant amount of time to figure out how best to move forward?

Aaron: It’s so hard to say, but people fall into three — a few different categories. One is there are still companies who are figuring out what they want to do with this, which for a company that’s been doing this for a while is a bit surprising — but still trying to figure out, you know, what are we allowed to do legally? Are we comfortable with this? What is going to be our data policy? How do we actually go about doing it? What’s the right tool? So they’re still kind of in that hesitancy and paralysis version of things.

The second group is, hopefully becoming less common, but is sort of let’s get everybody a Claude license or ChatGPT license and just let them use it for whatever they decide. They’ll figure it out.

And then the third is sort of like the more thoughtful approach. When I talk to organizations that are in that second group, that have kind of just decided to buy licenses for everyone and let it rip, a lot of this does resonate. So I am hopeful that it will start to be a conversation that becomes more common as people move more into these tools becoming more commonplace in the workplace.

And it definitely resonates with people who haven’t quite started their journey, because it’s very easy for me to say, here are the things that I wish that we knew when we first started that you can design for, so that you don’t get to a point where you do need to start to build back those bends in the river. You can take that into account from the very beginning, which is a huge advantage.

Jeff: Does the tech world have to set the standard for this? In other words, because you’re dealing with it really at the cutting edge, does the opportunity exist to set the standard for other businesses, whether it’s marketing, law, advertising, whatever it might be?

Aaron: It’s a good question, what it would mean to set the standard. I hope part of the reason that I wanted, particularly, to write the second piece was just to raise it as a topic of conversation, because I do think there is a right way to do this. And it is with humans kind of as the first priority. Whether or not we set the standard for it or not, I don’t know if I’m that optimistic, but I certainly hope to raise the conversation with as many people as possible. But we will see.

Jeff: Aaron Horwath has two pieces in Noema: Why Is Everyone in Tech So Sad? and Now Entering the Era of Blegh. Aaron, I thank you so much for spending time with us.

Aaron: Thanks so much for having me. It was amazing.

Jeff: Thank you. And thank you for listening and joining us here on the WhoWhatWhy podcast. I hope you join us next week for another WhoWhatWhy podcast. I’m Jeff Schechtman. If you liked this podcast, please feel free to share and help others find it by rating and reviewing it on iTunes. You can also support this podcast and all the work we do by going to whowhatwhy.org/donate.