Are AI agents coming for your job? In this episode, Brad Owens and Jennifer Owens explore the history of automation, from the Industrial Revolution to the rise of digital labor and AI. What makes today’s AI revolution different? And how can businesses and workers adapt?
Key topics discussed:
- The history of automation and labor shifts
- AI’s role in today’s workforce
- The difference between assistive AI and fully autonomous AI
- Real-world examples of AI in business
- The future of work and digital labor
AI is a tool—just like past technological advances—but how we integrate it into the workforce will define the future of jobs.
Links & Resources:
Research Paper on Automation & Labor Markets: https://www.nber.org/papers/w23285
Watch the full episode:
AI Generated Full Transcript:
Brad Owens (00:00)
my god Jenny, everyone’s gonna lose their jobs to AI agents?
Jennifer Owens (00:04)
People have been worried that they’re gonna lose their jobs to automation since the industrial revolution. What is different about today?
Hi, everybody, and welcome to the Digital Labor Lab, where we are exploring the future of work one experiment at a time. I’m your host, Jenny Owens.
Brad Owens (00:27)
And I’m Brad Owens and on today’s episode, we’re going to talk about the history of the future of work, kind of a back to the future kind of thing. Mostly what we need to dig into here is why is this such a big deal now? Why is everyone talking about this thing? But haven’t these sort of shifts happened in the past before? So Jenny, take us all the way back, back in time. Where do you feel like this sort of shift has already happened before?
Jennifer Owens (00:57)
since the dawn of time, human beings have labored on local agricultural-based units, know,
Brad Owens (00:58)
You
Jennifer Owens (01:04)
in a small community, family members, multi-generational. And then the Industrial Revolution happened, right? So a couple of things that I want to call out. First of all, our discussion here is pretty specific to the labor market in the United States. So sorry, European listeners, I promise we’re going to get to you. But today we’re going to talk about close to home stuff. So the United States started off as a primarily agricultural-based
economy heavily dependent actually on unpaid labor in the form of slavery. The industrial revolution happened, right? We saw a vast migration of people out of the farms and into the cities where we started to see some condensed places to perform labor and a real switch from subsistence farming to wage-based labor, which is primarily the lens that we used when we’re thinking about digital labor, although we’re gonna talk about that. So the industrial revolution happened.
Everybody got really excited about like cotton and you know, like fabrics, stuff you can make in factories. And then I think the next real revolution was when we started to see some automation and robotics in our manufacturing. So I’m thinking, please, yes.
Brad Owens (02:07)
hang out before we move off of this. Do you mind if I dig into this for just a second?
So what was it about the industrial revolution that really started making work different? What do feel like we really had that we didn’t have before?
Jennifer Owens (02:23)
I think that the main change for me is scale. So instead of an artisan working and producing, for example, articles of clothing in a cottage industry, right, so you might have a few different farms and artisans working, somebody might be tending the sheep, somebody else might be spinning the wool, somebody else might be fabricating stuff from that wool.
Instead, we’re seeing massive increases in scale. So we’re buying sheep at scale. We’re processing wool at scale and we’re producing clothing for purchase, not on a bespoke like, I know my kid’s going to outgrow their sweater, so I got to make a new one. But we’re really producing things for purchase also at scale. That’s okay. Take me, take me on a journey.
Brad Owens (03:03)
So I’m going lead you down a path here.
So what made that scale possible?
Jennifer Owens (03:14)
I we’ve kind of brought ourselves back into a loop, right? Because what was driving the industrialization was the ability to produce goods at scale. And I mean, because of factories, right? Like we’ve mechanized and automated part of that labor. So to go back to my example, right? Of the wool and the sweater, we’re no longer beholden to manual knitting, right? We’ve got machines that are capable of knitting and they’re capable of knitting much finer and different cloth than you can produce.
Brad Owens (03:20)
Why? Why do we have that ability?
Jennifer Owens (03:44)
So not only have we scaled up the production of a previously needed good, but we’ve also expanded the selection of goods that are available.
Brad Owens (03:52)
So you’re saying is, we had access to a tool that not only helped us do our labor better, but was far more efficient at doing so.
Jennifer Owens (04:06)
I see where you’re headed with this. Let’s follow this path. Tell me about your thoughts about how digitization and technical revolution can help us be more efficient in our labor and expand the options of what we’re able to produce.
Brad Owens (04:21)
So here is where I wanted to keep this thread throughout the entire episode. Because when we think of, let’s hang on to that industrial revolution. Now we started getting into the next sort of phases, which was we didn’t just have all these people sitting at these mills, at these refineries, at whatever it was that we were trying to revolutionize. We now had the ability to automate some of that work as well.
So now we had the ability to remove the human element from it. And let’s jump ahead a couple of years here, automotive manufacturing. The assembly line came along. We used to have a better way of working. But then what happened in the auto industry? We started getting robotics. We started getting things that could duplicate the exact same movements, the exact same way to make sure that we ended up with an end product that was what we had specified from the beginning.
But to be able to do that, we had to have the process down, locked in repeatable steps. And then we had one, I mean, let’s just take a robotic arm, right? Everyone knows that big yellow arm that sits in all the automotive factories. That one arm had a repeatable thing over and over again. So it was automated. It was robotics, but because it only had that one job, it was not intelligent.
Jennifer Owens (05:44)
So we’ve entered into kind of a, let’s call it a golden age of science fiction relationship
people and technology. The technology may be better than a human being because a robot doesn’t get, for example, repetitive stress injuries if you do perform all the proper maintenance. But the robot is only doing what the people tell it to do. We’ve got a very directive relationship between humans and their technology. So when I think about technology assist,
Maybe the second or third thing I think about is Clippy. Do you remember Clippy? Like back in the days when word processing on the computer was still like exciting and sexy and fun. And you could have like the little, thanks Microsoft by the way for your permission to use this. This is not true, we didn’t get permission. But I’m thinking about the little grammar and spelling assistant that would pop up when you were first typing your document. And you’d be like, Clippy, look, I am not doing a resume right now. Can you please stop? Go away.
Because although the idea was quite sound, the technology was not really there for humans to interact with Clippy in the way that we really want to be interacting with an assistant. also want to stop there and allow you to respond because there are three different directions I want to go.
Brad Owens (07:02)
Well, Clippy was a good first start, right? That was when we were trying to think of, how in a digital way could we assist someone in doing their job? And I keep using that word assist on purpose because it’s not doing the job. It’s assisting us in that job because we can dig into how we actually use AI ourselves at some point. But if you think about the majority of the ways that AI has been implemented up until this point, up until this next revolution that we’re going to get to.
AI has been used in such a way that it would assist us in doing something. It is a, we’re giving this thing a task by asking it a question of something that we need to do or giving it one specific job. So I always come back to it when people are like, AI is going to take over everything. I was like, yeah, how’s that auto correct working out.
Jennifer Owens (07:46)
Mm-hmm.
Brad Owens (07:54)
We’ve had autocorrect for a long time.
Jennifer Owens (07:55)
So I think it’s really interesting because I feel like in our current artificial intelligence landscape, the products that are available to the consumer are kind of chunky, right? So I’m thinking about our word processing metaphor here, right? I came of age before there were digital assistants, right? So I learned to use a paper dictionary and a thesaurus, and for example, an encyclopedia if I needed to look something up. So I am used to going to a specific reference.
for a spelling, a specific reference for another word choice, a specific reference for, wait, do I really know what happened in the student’s revolution of like 1863? Yeah, I don’t know, but the encyclopedia knows and I can go get that. This is kind of how I think of AI right now, right? We’ve got large language models, which are terrific at generating text. We have machine learning algorithms, which are capable of ingesting large amounts of data and deriving patterns and drawing conclusions from that.
At the consumer level, we don’t really have, and maybe I’m wrong here, but I have not yet seen an autonomous workflow that I would trust, even 60 % of the time.
Right, if I ask, you know, like chat GPT to say, okay, make me a meal plan, make me a grocery list from the meal plan. Okay, now go into my account and order these groceries for delivery. I don’t feel quite confident there yet. So.
Brad Owens (09:16)
yet. And
the key distinction that you said there was at the consumer level. These things are possible. It is completely possible. If you know a bit of deep technical knowledge of jumping into Python, or you have an exorbitant amount of money to spend on open AI’s new orchestrator or something like that, some of these things are possible.
Jennifer Owens (09:20)
Yes. Yeah, absolutely.
Mm-hmm. So today, right, we can use task-specific assistance. What about this is driving the conversation about what might be possible tomorrow? Let’s talk about autonomy. Let’s talk about how this might really shift our workforce. And then I want to take us back.
to that thing that you said earlier about we have a better way to work. As you were talking about automotive manufacturing and we’re talking about this, because I really want to probe that from a couple of different lenses. But first, let’s discuss the future.
Brad Owens (10:08)
So when we, when I think about my day job, what I have been able to.
Jennifer Owens (10:14)
Wait,
pause. For those who are listening to this episode first, what’s your day job, Brad?
Brad Owens (10:18)
So my day job, I am working with Salesforce space technology to help companies come up with a way to do their work. Better for lack of, for not going into a ton of depth. That’s what my day job is all about. We have a consulting organization that is able to help companies do work better with the Salesforce platform. And because it’s on Salesforce, we all know Mark Benioff’s feeling about the future of digital labor. They are driving hard into agentic AI.
Jennifer Owens (10:47)
Wait,
what if I’m learning about this for the very first time? What is Mark Benioff’s stance on agentic AI and digital labor?
Brad Owens (10:55)
like it. So Mark Benioff was on stage at the world economic forum in Davos, and he was talking to all the other leadership that was on stage and said at one point, Hey, are you all aware that we are the last leaders to lead only human labor? And one kind of pause and looked at him and he went into this detailed explanation about. We only currently as leaders and as business owners have had to lead human labor.
what he is picturing and honestly what the Workday CEO, what the Google CEO, everyone is starting to understand is that in the very near future, it is highly likely that we will have digital employees. What we’re referring to on this show as digital labor. That’s made possible because of what everyone is starting to talk about and, oh my God, it’s going to come take my job, agentic AI. That’s what we really need to dig in.
Jennifer Owens (11:53)
So what’s interesting is that as we were conceiving and really refining the idea for this podcast, I started to do a deep dive and do some research. And one of the places where I went was actually golden age of science fiction. I went back to Asimov and our three laws of robotics. I started to think about the relationship of humans and their technology, because this is something that is really fascinating to me. When we approach these kind of.
changes from a place of fear, right? It feels like AI is coming to take my job. When in reality, I feel like a hybrid human digital workforce looks a lot like our space program sending rovers to Mars and teaching them to sing happy birthday to themselves. like, I know, right? Like, like, so I think it’s really interesting, right? The desire of human beings
Brad Owens (12:42)
Just, bah, all right.
Jennifer Owens (12:48)
to kind of have dominion over something, right? That very golden age sci-fi like the human is in charge of you robot, you robot do what I tell you to do versus humans desire to make pets out of stuff. And I think the future of digital labor really will succeed when we have labor that we can feel friendly about, that is truly an assistant, that we feel is working hand in hand with us. I have a lot of other thoughts on iRobot that I will…
leave out for the sake of time. But if anybody wants to talk about Asimov with me, I’d love to. I’m curious, though, because we’re talking a lot about assistive artificial intelligence. And we’re talking about the future of digital labor as being truly agentic, so autonomous, having that agency to generate a, you to respond to a prompt and then to go and take action on that. And it’s interesting, as I’m thinking about what are the impacts of this going to be on the labor force, right? So switching perspective,
from the perspective of the employer thinking about, how can I get more productivity out of the human capital that I have versus now I’m also the employee of an organization. How am I gonna experience this change as a member of the workforce? And this is really interesting to me because our show is called Digital Labor Lab, right? We wanna do research, we wanna do experiments, we wanna think about things. So I found this really interesting paper on automation and local labor markets from 2017.
This was in the National Bureau of Economic Research in 2017. The authors are Asimoglu and Rastrepro. And I thought that this was really fascinating because they looked at the impact of robots and automation on local labor markets. if you’re, the paper is 91 pages long, it’s a fantastic read. really do recommend it.
link down below. But the short term takeaway is when the robots are competing with human labor on various tasks, those robots, then the presence of robots in the labor market reduces employment and it reduces wages. Right. So when humans and robots are competing, absolutely humans are going to lose. Right. robots don’t get repetitive stress injuries.
This is interesting to me because I want to think less about a labor market in which humans and agentic AI are competing for roles and more about a labor market in which humans and agentic AI are collaborating. So how can we use these to remove some of the stresses of our workforce? How can we use these to do the jobs that people aren’t great at at scale?
or individually? How can we use agentic AI to do what AI does best and what people don’t do great and free up people to do the things that people do really well? I keep thinking about the tweet that keeps circulating, right? I don’t want AI to do the art so that I can do more dishes and laundry. I want AI to do the dishes and laundry so that I can do more art. That’s what I love. I want to see artificial intelligence do the stuff that people either don’t want to do or aren’t great at doing to free us up to be the most human we possibly can be.
Brad Owens (15:50)
So I thinking about an analogy for this today. And if my grandparents, if they were still around, if they came to me asking, Hey, what is this whole agentic AI thing all about? There’s the fear side of things of, my God, this is going to come take my job, which we can dig into because it may not right now. It is a tool. It is a tool to be used and.
If I were to able to explain this in a very simple way to someone that may not understand what this is capable of and to dig past their fear about this, I’m glad that they’re calling this agentic AI. think on purpose, that was a fantastic call because if I were to think of something as an agent that I would interact with on a daily basis or something that I have experienced with, I think back to a travel agent. And I want to talk a little bit about.
the experience of working with a travel agent. did it once. We had a recent trip that we wanted to go on. All we knew was here’s where we want to go, the types of experiences that we want to have, and here’s roughly when we can do that. That’s all we said. Now the agent, the travel agent went off, did all of the research, came up with the entire plan.
booked all of the vacation locations, helped us with understanding what travel arrangements may actually get there on time, made sure that we had experiences that fit what we wanted to do while we were there. We just sat back and had an amazing time. When we think about that sort of interaction with agentic AI now, we’re essentially talking about that same sort of experience, but with our work. So think about the types of things in your work that may be annoying the ever loving crap out of you.
And how would that feel to offload that to someone that could take that for you?
Jennifer Owens (17:46)
So there’s two things that are interesting to me. One is that you spoke about our recent experience with the travel agent and I agree it was a wonderful experience. We had one 15 minute phone call in which we prompted the agent with the kind of things that we wanted out of the trip. And the second thing that was interesting to me is that we had that experience in 2022 at a time when we were like when when travel agency really has has kind of suffered right from the rise in Yelp and TripAdvisor and all the other stuff.
There’s a lot that you can successfully do yourself, but for a trip of this magnitude, right? This was really kind of a once in a lifetime trip for us. For a trip of this magnitude, we didn’t want to do it ourselves. We wanted an expert. So we went to the expert.
The second thing that this really brings to mind though is, you’re talking about our work and I wanna think about work not just as something that I’m doing in exchange for a paycheck, but also the other things that are enriching my life. So, I volunteer in a couple of different areas, right? But one of the things that often sticks out to me is the challenge of orchestrating volunteer labor because people are available when they’re available. Some people are available Wednesdays from like 10 to two, but not if the Wednesday is a prime number.
And like untangling all of that is a fantastic job for artificial intelligence. And it is a pain in the butt for a human being to figure out, okay, I have these six people, here’s their availabilities, I need to staff this location for these hours, go. I think it’s interesting to think about the gains that can be made from the use of digital labor in places that aren’t necessarily the exchange of labor for cash.
Brad Owens (19:26)
So where I want to bring this back to and help everyone to have a solid grounding in is there is a lot to bring this completely back to where Jenny started this conversation. There is a lot about this that is still possible, but the majority of it is still science fiction. I can tell you from seeing it in the actual wild,
We’re not at the point yet where this is going to produce some kind of mass layoff because of AI. It’s just not there
Jennifer Owens (19:55)
Remember that eating disorder association that had laid off all of their phone line people because they had like an AI thing that was gonna do it? And then it turns out that their AI tool was giving fantastically irresponsible advice, so they hired their people back.
Brad Owens (20:08)
So it’s not there. It’s absolutely the speed of change is coming that it could get to the point of starting to replace a large portion of the workforce or where we want to keep the focus of this right now because it’s what’s most applicable to the majority of business owners out there is the assistive technology that is out there is getting to be pretty tremendous.
And helping people to do their job at a level that was unheard of before, but just like the industrial revolution, this is a tool. This is a tool that allows you to make the best use of it for you and your business. I don’t think that the majority and Jenny, I’m curious to get your take. I don’t think the majority of businesses are at that point where they need to start thinking about how am going to replace my people with digital?
Jennifer Owens (21:00)
I think the other large unknown in all of this, and this is something that I would love to dive into on a future episode, is how does the automation of the workforce, the agentification of the workforce, drive your consumer product or your service that you’re providing? We just talked about that example with the eating disorder line where the output was unacceptable. And so we ended up seeing a resurgence of human labor. I’m curious to know if I’m a
If I’m a, I keep thinking about like the robot nail painting things that you see at like Vegas or sometimes in like the fancy airports. If I’m thinking about that, where will the consumer start to make a choice that’s different? What of the agentic workforce is acceptable to a consumer or a purchaser of services? And where do we want to interact with a human? That’s interesting to me. And the other thing is that this fear that AI is taking our jobs,
I think that’s interesting. I think we should lean into that. I think we should use that as a lens to explore how we feel about our work. What about our work is truly uniquely us, uniquely human, and what truly can be automated without sacrificing quality, without sacrificing the care? I think feeling that fear and using it as a lens to explore is the direction that’s going to be most productive.
Brad Owens (22:22)
And that’s why we have started the digital labor lab and why we hope this would be exciting for you all to listen to as well. Because these, at this point, are all just experiments. Even the big headline grabbing things of Salesforce saved 50 % of their caseload because they were able to take all of their customer service and move that to a bot. That’s right now just an experiment for them. They’ll even say it themselves. It’s just an experiment. So that’s why this is a digital labor.
want to give you the freedom to be able to experiment with these things, to find what works best for you, for your business, to be able to help you grow and to take advantage of what’s out
So based on where you’re listening or watching this episode, hit that subscribe button for us. Make sure you follow along with this and all the other episodes that are going to come. Jenny, where can they find us?
Jennifer Owens (23:08)
You can find us at digitallabourlab.com. You can also find our episodes and bits and clips on many social media platforms, Blue Sky, LinkedIn, all sorts of places. We would love to have you engage in the conversation there as well.
Brad Owens (23:21)
If you search digital labor lab on your social media of choice, odds are we’re going to be there. So until next time, I’m Brad Owens. We’ll see you.
Jennifer Owens (23:27)
I’m Jenny Owens.