AI Eliminates Tasks, Not Jobs
With guest Greg Godbout, Eleven Canterbury consultant and former CTO at the FDA, and host Dan Martin, Eleven Canterbury Program and Relationship Manager
Summary
AI isn’t coming for your job. It’s coming for your tasks.
That distinction will separate the organizations and professionals who thrive from those who fall behind.
Greg Godbout, former Chief Technology Officer at the EPA and former Chief Digital Officer at the FDA, joins Dan Martin to explain why every major technology revolution has eliminated tasks, not human work. The real opportunity isn’t to replace expertise. It’s to amplify it.
Their conversation explores why AI will create new roles rather than eliminate existing ones, why today’s education system may be preparing students for yesterday’s workplace, and how organizations that embrace AI will solve problems that have long been out of reach.
They also discuss:
- Why AI is shifting work from repetitive tasks to judgment, orchestration, and decision-making
- How AI assistants can dramatically increase the impact of knowledge workers
- Why businesses may need to teach AI skills faster than schools and universities
- How AI can modernize decades-old legacy systems that organizations have been afraid to touch
- Why freeing technology budgets from maintaining outdated systems could unlock unprecedented innovation
- How AI and human expertise together will tackle problems that were previously too complex, too expensive, or too risky to solve
The biggest winners in the AI era won’t necessarily be the biggest companies. They’ll be the organizations that rethink how people work, embrace new tools, and use AI to achieve outcomes once thought impossible.
The biggest question isn’t whether AI will change how we work. It already has. The real question is what becomes possible when people stop protecting old ways of working and start solving problems they couldn’t solve before.
Transcript
Dan Martin: Today we’re talking about AI, its effect on the job market, how jobs will change, and how companies will succeed by embracing AI.
I’m lucky to be talking with Greg Godbout. Greg is the former Chief Technology Officer and Chief Digital Officer at the EPA. He is the Chief Growth Officer at Fearless, a Presidential Innovation Fellow, and a federal employee top 100. An incredible series of accomplishments in a wide variety of fields.
AI is such a transformative technology. Every time in the past where we’ve looked at transformations, the success goes to the companies that are able to take advantage of the transformation. It’s rarely the company that was the biggest in a particular area beforehand. Hilton doesn’t own Airbnb. Kodak didn’t prosper with the digital camera. Yellow Cab didn’t invent Uber.
What are the things that are going to make people successful in this new era? People are worried about the demise of white-collar jobs.
Greg Godbout: First thing I would say is that on all these major technology shifts throughout history, there are two common denominators.
The one you mentioned, absolutely, it’s not the leading industry companies that invent the disruption. They usually are the ones who are pained by the disruption, and that has something to do with them not thinking about or innovating on that disruption. But a big aspect of that is that at every major technological shift, we’ve eliminated tasks that humans used to do, but we haven’t eliminated work for humans.
AI is no different. We’re all going to be doing the same thing, but much faster. And when I say doing the same thing, I think a lot of people – and there’s a lot of worries, and I understand it, a lot of people are worried that, “Well, if I’m not going to do my task anymore, then am I going to lose my job?”
When really, probably when they were hired, or even early when they were hired, they had a thought of – My job is to achieve some sort of outcome. And you really shouldn’t dwell on the tasks. They’re not nearly as important as achieving that outcome. And if you could have automation and AI and tools to take away a lot of tasks, but still achieve that outcome with higher quality, that is going to drive the day.
So, when you talk about what’s happening, say like white-collar jobs or things like that, everybody’s role is going to be affected by AI. It’s not just white-collar and blue-collar. I think with LLMs, white-collar workers, for the first time, are thinking, “Oh, these are the robots that are coming to replace our job,” versus, say, a blue-collar worker who’s seen this in factories already. Or in warehouses.
I think the reality is the jobs just shift. Every technology revolution creates more valuable jobs than it eliminates. It doesn’t replace expertise. It really amplifies it. So the outcome will not be replacing workers, but allowing every worker to accomplish dramatically more and achieve dramatically more above what they used to be able to achieve with the knowledge that they had before.
When they can have their own assistant, that is an AI assistant, enhancing their job, they’re going to move into a different space. For years people have been talking about the info worker. Humans are about to become the info worker. So, when you think about white-collar jobs, I think in many ways it’s going to create more white-collar jobs because we’re going to have info workers and management and people orchestrating these tools along with other people.
Dan Martin: So, it’s different kinds of tasks and orchestrating. Do you think people will be prepared for that? Are we educating people to do that kind of job, or do they grow into it?
Greg Godbout: I think, with AI, the transition’s going to be so fast. There will be a sense of people feeling whiplash. Of like, “What do you mean I need different skills?” So I think that’s a reality, and that’s a reality of any of this change.
Unfortunately, I don’t think our education systems are really, and this is true at every technological advance, the education systems are in a bygone era. They’re protecting the old technology, and this is happening again. Just this week there was news of a professor catching people cheating using AI on one of his tests. I don’t doubt that, based on the rules of the test, it was probably cheating. What I find interesting is why we are not allowing everyone to use AI on all tests and all papers, and why we aren’t changing the tests and the papers’ expectations? Because when that group of students gets into the workforce, there isn’t going to be a company that’s like, “Hey, we’d like to hire you, but don’t use AI.” They’re not going to consider that cheating; they’re going to be like, “No, in order for you to be a productive worker, you better come in with all these AI skills.”
So, I think that pain point is going to be pushed on the businesses. The businesses are going to have to get people trained up faster, because unfortunately our educational organizations are much slower to adapt. You can see it at high school and college levels. I talk to high schools all the time about this. They still primarily all think of it as cheating. But I know of private high schools and some public high schools that have now changed their whole curriculum based on it, and those are the early adopters, the innovators that are like, “No, this is the new workforce tool that we need to teach our students to be prepared for.”
And frankly, target those schools to hire from. Otherwise, businesses are going to have to do a lot of that upfront.
Dan Martin: In the education system, we’re teaching people not to use the tools they’ll use when they go out and get a job.
Greg Godbout: Right. Can you imagine? Let’s take something else. Let’s say we were going to go to war at the invention of the musket. And you had a war college in your country, and it was like, “We’re not going to use the musket. We’re going to stick with arrows, and never move to a rifle or gun.”
You’re literally not teaching the tool that you’re going to face, and they’re going to compete with people who have become AI-skilled, AI-enhanced workers. It’s unfortunate. It’s going to put students at a disadvantage in those education systems.
Dan Martin: This is what happens to companies when there’s a new innovation; they worry about losing or cannibalizing what’s been making money for them.
For people, when their job is under threat from what they perceive as AI, white-collar jobs, they worry that AI will do the same thing a whole lot faster and my company won’t need me. That’s short-range thinking, isn’t it?
Greg Godbout: I think change is scary, and I think for a lot of people it’s hard to think about: how will my job change, and how will I like it? But inevitably, what happens in each case is that the more repetitive tasks are taken over by AI. The tasks made jobs a little less fun. You didn’t use your critical thinking as much. It just became habitual. Those are the tasks that will be taken over from an efficiency standpoint.
But in addition, there will be more access to knowledge and data for those people working, that they’re about to explore. They’re going to be the coordinators, they’re going to be the thinkers, they’re going to be the decision-makers who actually orchestrate this whole thing.
AI is amazing, and it is equally dumb at the same time. It needs to be well managed, and it needs human judgment to reach a higher level. That expertise is not just you take a human, you say, “Okay, I have a domain expertise in, say, sales or something, that’s my job.” So that person becomes very valuable in helping to build out the new sales workflow, right? And then they can automate things; they can talk about things and automate things they never imagined before, and then they can learn more information. That information gets fed to that human, and then they learn faster.
We’ve computerized data for decades. Now we’re computerizing knowledge, and who benefits from that knowledge but the workers themselves. They’ll exponentially learn from that, and then be able to educate those systems. So it just creates a whole new type of skilled workers constantly weaving and orchestrating, achieving the same outcomes, but just giving up on some old tasks and taking on new tasks that frankly are more rewarding.
Dan Martin: What you’ve described is that we’re using the technology and AI in areas to improve efficiency, but using the whole system to improve effectiveness in what we do. Not only doing it more efficiently, but more effectively. So, the winning strategy would be doing far more work with the same or more people than what you were doing before, right?
Greg Godbout: Yes. And achieving greater outcomes.
I’ll give one quick example. This was with a company I was doing work with in Denmark. They had built a system for the group inside the government that managed marriages. So essentially, when people were getting divorced, they went to this group. All they did was manually process divorces, separations, and things like that. And eventually they digitized and automated their processes and took away the repetitive tasks. It saved a bunch of time. The workers said, “Well, let’s meet in the afternoon and talk about new ways to do this.”
They had gone through the digital transformation of everything. You know, moving from analog to software, to all this stuff. And then they got to this spot where they’re like, “You know what? We have so much data, and we’re now connecting to these other departments; let’s figure out the money for these situations. Do they have children, do they have all these things?” They’re like, “Could we figure out what marriages or what people are in crisis? Could we become the healthy marriage office?”
So they started thinking differently about the work that they were experts at. When you got them out of the routine, they, on their own, came back and started asking for – they didn’t know what to call it, but they were asking for AI and advanced analytics to help them see and discover things they couldn’t within all of this data.
People, we all, naturally evolve that way.
Dan Martin: As you talk about that, I think about some of the kind of intractable problems that might be able to be attacked with AI and human beings.
When I was a young student at Georgia Tech, the largest computer application we had was the very important season football ticket application program, written in COBOL, which they probably are still using. Banks that I’ve worked with have huge backlogs of things written in COBOL and even assembler language in different things that are kind of intractable for people to attack. But is there something that, in combination with an AI agent and people, you could approach some of these things? Is that a good use as well?
Greg Godbout: I’ve been doing federal government and public sector contracting for years. Our CMS runs Medicaid and Medicare. The largest healthcare system in the world runs on three COBOL applications. The whole thing is built on these three applications, right? Two of them are 30-plus years old, and one is 40-plus years old. The 40-plus-year-old one, they don’t touch anymore and upgrade. They’re afraid to. It just works it, right? And you can imagine, this is massive. And you can imagine everyone working around this just basically holds their breath and is like, “Oh, please don’t break.” And every attempt to modernize these systems- and this is across- you mentioned financial institutions have this problem, but states have this problem. Massive systems that just became legacy systems written in some old technology that’s no longer supported by the people that created that technology, stopped supporting it decades ago.
Dan Martin: And the people that wrote it are retired or dying.
Greg Godbout: So then you’re sitting there, and you’re like, “Well, do we train new people in COBOL because we’re afraid to update these systems?”
Then along comes this technology that can take the knowledge of that system, not just the code. The code’s filled with bad code, right? It takes the knowledge of the system. It can create a whole new operations model and business model based on reading that code. It’s knowledgeable enough to do that.
Dan Martin: Well, now all of a sudden you can take the leap, and if you’re doing vibe coding and using tools like Codex or Claude Code or something, you can have a team build out the new application in weeks, not years or months. Or decades or centuries.
Greg Godbout: Right. I know. We’re talking about very long times here.
Then that becomes: what we’re really doing is we are now able to solve problems that were too difficult for decades. We can now solve them rapidly, and it’s not just, oh, it’s a better way to write software, or it’s more efficient. That’s true. But the number of problems we can solve now is exponentially high, and that’s going to drive more work, and that’s ultimately what drives the workforce. The demand for more people to do this. You’re not going to just have machines automate the leap to a new mainframe system. You’re going to have them do approximately 80% of the work, but you’re going to need a whole new host of humans to oversee it, security-check the code, and evaluate and orchestrate it.
So there’s a whole universe of infinite problems that we haven’t touched that we can now start working on. It’s going to create a demand for work that drives all the new jobs.
Dan Martin: A company that I worked for had an ancient technology just like that. They were afraid to move it, and they built a new computer room around this
We’ve got opportunities where you can do things that, not only you couldn’t attack before, but if you did attack before, there were significant risks to the people in charge of doing it.
Greg Godbout: Well, and think about the cost of that. Budget-wise, most federal agencies, most public sector agencies, and your largest organizations that have these large systems, in general, it’s usually about 80 to 95% of their budgets is just taking care of these old systems.
Dan Martin: It’s precisely right. I worked with several large banks, and you would talk to people about doing new things, and you’d hear, “95% of our budget is just keeping it running.” In the environment where you have new regulations, you have new rules, you have new products, just keeping it going.
Greg Godbout: So, then you free up the budget, and somebody might say, “Well, hey, let’s save some money,” which some people are going to try. And other people are going to say, “Hey, let’s do all the things we’ve always wanted to do now. Let’s spend 50% of our budget on new innovations and new leaps and bounds.”
And you’re going to see an explosion of productivity, of creativeness, and innovation. The companies that embrace it, their sales are going to skyrocket. And the organizations and mission areas that embrace it, their effectiveness and their ability to address the mission is going to not only grow, but they’re going to be able to solve the problems they never could before because they didn’t have the resources and time to fix this obvious problem.
Dan Martin: What an exciting future we have. Solving problems we couldn’t solve before.
Greg Godbout: It is.
Dan Martin: I’ve really enjoyed the discussion, Greg. Thank you for sharing your opinions.
Greg Godbout: Thank you so much.
