Automation and Imagination: The Futures of Work with AI

From fears of mass unemployment to the power of "boring plumbing": What five expert conversations taught us about AI’s impact on work

Created with ChatGPT 5

After five interviews with researchers and practitioners, we sat down to compare notes on how AI is actually reshaping work. In our conversation, we covered fear of job losses and techno-human collaboration and, why “boring plumbing” may matter more than general chatbots, what leaders should (and should not) do, and how regulation and geography shape our use of AI at work.

Markus: Over the past weeks, we talked to five smart people. I’m curious, Daniel – what stuck with you most? Is there one insight or one conversation that keeps echoing?

Daniel: The theme of our first season was AI’s impact on the labor market and my key take away is that even though AI won’t make us all unemployed, it may feel like it. The potential shift of certain white-collar jobs to other countries was one of the predictions that Carl Benedikt Frey made in our conversation.

People read about mass layoffs in the media and about AI’s potential to boost productivity on the next page and they connect the dots. But from our interviews, I’d say we’re seeing correlation more than causation. This creates a challenge for all of us: How can we adopt AI without creating more anxiety?

Markus: I had the same thought. Many cuts we’re seeing in the German industry can’t be blamed on AI. But the shrinking number of entry-level white-collar jobs is closer to the AI zone, as Carl also noted. And his point about business work being easier to relocate is unsettling. It hints that in professions like consulting or law, Western markets may see fewer domestic jobs. It brought back the “the world is flat” debate started by Thomas Friedman who argued that many tasks could already be done in India, for example. Maybe that wave is returning, now supercharged by AI. What do you think?

Daniel: To me, this is a trend of globalization generally and not tied to one technology specifically. And graduates have always been more exposed to risks in the labor market – either from weak economies or automation of routine tasks. When I started my career, someone had to clip physical newspapers for the daily press review. That job’s gone now. Graduates still face headwinds – but they also enter the job market better equipped because they grew up as AI natives.

Markus: True. Still, the old calculus – hire juniors to do low-value work while they learn – may be breaking. If AI handles the grunt work – research, slides, first drafts – does the “train them on simple tasks” model still pay off?

Daniel: That’s an open question, I believe. We touched on this with Hans Rusinek when he told us how his students are preparing for the job market. And the range is wide: there’s anxiety about unemployment on the one hand and the possibility that, with the help of AI, graduates can become advisors to CEOs right out of university.

Markus: And skip joining big organizations altogether. Hans hears more students asking, “why join a giant firm? I can go solo.” It’s the old indie dream, and now we have the tools: a laptop plus a swarm of AI copilots and expert agents.

Daniel: Which makes the question of how to implement AI so important. I like GitHub’s playbook approach very much: It reduces the fear of being replaced by treating AI as a normal tool, like any software – although a far more flexible piece of software. If you frame adoption that way, you can still extract tons of value from early-career hires by just shifting what they do.

Markus: Lufthansa’s Cindy Richter said the same for marketing and comms: AI requires new skills and creates new possibilities. Some are simple like animating comics for kids with AI and some more strategic – helping employees act as credible brand ambassadors. It didn’t sound like “fewer people” but more like “different work.”

Daniel: That was exactly Carl’s point: if we only see AI as automation, we miss the upside. Yes, AI automates repetitive tasks. But the real way to advance our society is creating new things with AI – jobs we can’t fully imagine yet. We’re still far from superintelligence and we will continue to need humans to set goals, give instructions, supervise outputs, and thus create economic value.

Markus: The shiny “new industries” story is inspiring … and a bit hand-wavy. The flip side is what Jonas Andrulis emphasized: the “boring plumbing”. Not chatbots on the front, but re-wiring back-office processes, data, and workflows. That’s where real productivity hides.

Daniel: And “boring” can also be transformative as we have seen with Lufthansa and GitHub. We’re still in an exploratory phase where we are discovering the value-add of AI.

Markus: That requires leaders to allow exploration – and that is much harder than it sounds.

Daniel: Leadership certainly plays a big role. There are two very different lenses on how companies view AI. The first is efficiency or cutting costs in the old system. That’s surely legitimate, but limited. The second is exploration: Challenging your own business model before someone else does. That demands a learning culture, tolerance for controlled failure, and executive air cover. I believe that this is the more important mindset, because we’re entering an age of discovery that needs inventors, not managers.

Markus: Regulation complicates the picture. You’re our policy brain – what did you make of Jonas’ warning about over-regulation? He said the real risk in Europe is moving too slowly.

Daniel: He flagged regulatory capture – the risk that complex rules entrench incumbents who can afford compliance. And this is an important warning. Too often small firms hesitate to adopt AI because they fear compliance burden. Luckily, Europe is currently considering adjusting the rules so that they don’t stifle innovation; that’s an important discussion to have.

Markus: Politicians love to say “a distinct European way rooted in our values”. Fine – but does that quietly mean “more rules”? You posted about this recently; it’s a tiny example but does it stand for something bigger?

Daniel: We should avoid repeating the mistakes we made on cloud adoption where we spent years lamenting the lack of a European cloud and then under-utilized the tech that exists. When you’re catching up to a new technology, a heavily involved state can help close gaps – Carl makes that point in his new book. But once you hit the frontier, innovation is inherently decentralized. Europe still wrestles with the gear shift between “use rules to ensure safety” and “loosen rules to unlock invention.” And that is arguably a very hard problem to solve.

Markus: We also revisited the “flat vs. spiky” world debate – Thomas Friedman vs. Richard Florida. Carl emphasized that big cities remain the engine rooms of innovation, including AI.

Daniel: As a Richard Florida fan, I’m glad that the creative city thesis is still fitting. It underscores the role of urban planning. During the pandemic, there were many takes about the alleged demise of San Francisco. But most of the AI labs are located there – and this is certainly no coincidence.

But let’s go back to the core question: does AI destroy work? You argued in your last book that AI could make work more meaningful – less drudge, more purpose. Would you still say that’s the case?

Markus: After our conversations, I would double down on that. Hans worried about our always-on, distraction-heavy workplaces killing the “incubation” phase creativity needs. He hopes AI can strip out admin and alignment to give us focus time. That’s also what tech vendors promise, although reality is arguably lagging behind. Maybe that is because as Jonas said, building AI is not trivial. You must fix data, break silos, and redesign processes. That’s work. But the payoff is real.

Daniel: We only touched on the issue of loneliness. If your closest “colleague” becomes an AI, could that isolate you? Or does it just mean culture has to be built more intentionally?

Markus: I’m wary of over-engineering culture. It emerges. But yes, the mix changes. Many power users already run “panels” of AI agents to stress-test ideas from multiple angles – the kind of perspective-gathering we used to do by putting ten humans in a room. Will that make us lonelier? Not sure. It will change how collaboration feels.

Daniel: And let’s be honest: we always dreaded big opinion-gathering meetings, no? I’d rather walk into a meeting with a tight, well-sourced brief from my AI assistants – and spend human time aligning on the truly important questions. AI shouldn’t take those away from us.

Markus: That’s basically what good leaders already did: collect perspectives, invite dissent, then decide. Now we all need that leadership muscle. AI can prep the ground and reduce alignment thrash, but judgment remains human.

Daniel: Looking ahead, what is your thesis on how AI and work will develop?

Markus: First, there’s real money in automation and streamlining. Products that actually save time will land. I expect many of us to reclaim 20–30% of a week from tedious tasks. That’s a once-in-a-generation chance to redesign work. Second, I am more skeptical than Carl and you about the potential of AI helping us to develop new industries. It takes a specific mindset and willingness to change the world which I don’t think we have here.

Daniel: I’m watching for a “Henry Ford moment”. Ford didn’t invent electricity or the conveyer belt; he recombined existing technologies into a new production system – and unlocked massive value. Similarly, I’m waiting for firms that don’t just do the same things faster, but do different things altogether. Once that switch flips, we have arrived in the age of AI.

Markus: Or – hear me out – it may be less a flip and more a fade. That “AI as Normal Technology” essay convinced me. I don’t think there’s going to be an AI apocalypse or a single “AGI day.” Just a steady drumbeat of small improvements in our work life. In two or three years we’ll look up and think, “It’s wild that we once met on video call just to align who does what until when”.

---

We want to hear from you: What were your take aways from the first season of our newsletter? And what would you like to discuss with us in season two?

---

Season 1: AI and the Labor Market | Episode 1: The Future of Work, In Progress | Episode 2: Carl Benedikt Frey: “Professionals are not prepared for the coming changes” | Episode 3: Jonas Andrulis: “Digitize the state! That’s the foundation we all stand on” | Episode 4: Cindy Richter: “AI creates roles that didn’t even exist before” | Episode 5: Matt Nigh: Matt Nigh: “You can’t force AI from the top down – it needs energy from the ground up” | Episode 6: Hans Rusinek: “When all companies use AI, the human factor becomes decisive”