What does it mean to be human in the age of AI?

Franz has a unique background combining AI, journalism, and a PhD in the philosophy of science. He currently leads Innovation & AI at Blinkist, where we worked together on a new AI product. Last year he published a book on using AI to improve your life, and now spends a lot of time thinking about what it means to be human in the age of AI. In this interview he shares some of his reflections, which are particularly relevant as all of our jobs seem to be fundamentally changing.

You and your co-author Dirk von Gehlen recently published a book called “Wie KI dein Leben besser macht.” What’s the main takeaway?

Our book emerged at a moment when a lot of people – especially those of us coming from the humanities – were still trying to make sense of what generative AI actually means. And I think something gets overlooked in a lot of AI discourse: there’s a kind of quiet affront that comes with it. When you realize that capabilities you thought were distinctly human – writing, reasoning, expressing ideas – can now be replicated by a machine, at least to a significant degree, that stings a little. It raises uncomfortable questions: Are we going to forget how to think? Will we become intellectually lazy? What does this actually do to us?

What Dirk and I share is a posture of openness without naivety. Dirk is a journalist at Süddeutsche Zeitung who has spent years thinking about how the internet reshapes our thinking and culture. We’re not hype merchants, but we’re also not catastrophists. The book was written for people who don’t work in tech – reasonably curious, open-minded people – and our basic message was: you won’t be able to avoid this topic, so you might as well engage with it thoughtfully. AI will likely be one of the defining forces of this century. The question isn’t whether it affects you. It’s whether you shape how it does.

I framed this under the term digital humanism – the idea that good technology use carries genuine enlightenment value. That there’s something worth fighting for in making sure people aren’t just passively overrun by these changes.

The idea that got the strongest reaction was what we called the Socratic Machine. The thought experiment goes like this: what if you configured an AI assistant to disagree with you as thoroughly as possible? To push back on everything? Combined with the ability to have actual voice conversations with AI, this opens up something genuinely new – or rather, genuinely old. For centuries, the dialogic mode of thinking – working through ideas by arguing them out loud with someone who pushes back – was considered the highest form of inquiry. AI might finally make that widely accessible again.

In your book you talk about using AI in our lives. How do we identify what to offload to AI versus what should remain “human”?

There’s a distinction outlined by Steven Shaw and Gideon Nave of the University of Pennsylvania that I find helpful: the difference between cognitive offloading and cognitive surrender. Cognitive offloading is what you do when you use a calculator – you shift a discrete task to a tool, and that’s fine. Cognitive surrender is something else: it’s when you stop exercising your own judgment entirely and simply adopt the AI’s output as your own.

Based on this distinction, my answer is this: offload the things that are essentially mechanical – formatting, first-draft research, scheduling, repetitive writing tasks. But protect the things where the struggle itself is the point. Ezra Klein put it well in a recent New York Times essay: the work of chiseling a vague thought into something publishable isn’t just a means to an end. It is the thinking. If you skip that process, you don’t just save time – you forfeit the development that would have happened inside it.

I think the clearest signal is: does this task require you to grow? If yes, be careful about outsourcing it too early. If it’s something you’ve already mastered and it’s just costing you time, offload it freely.

The deeper principle, which connects to what Dirk and I wrote about, is that the goal should be amplification, not replacement. AI should make you more yourself – sharper, better-informed, less bogged down in friction – not a pale imprint of a generated version of you.

What distinguishes those making the most of AI from those struggling to keep up?

I have to be careful here, because working in tech makes me susceptible to a particular bias. There’s a narrative that circulates constantly in this world that sorts everyone into two camps: those who are all-in on AI, and those who are resistant, with the clear implication that the second group are the losers of history. I don’t think that’s an honest picture.

Reality is much more granular. Some people are so busy playing with new tools that they never actually get anything done – procrastination dressed up as innovation is still procrastination. There are those who are genuinely curious and open but face real legal and organizational constraints. And there are others with entirely legitimate concerns – about corporate power concentration, about copyright, about what’s actually in these training datasets – that have nothing to do with being a Luddite. The big questions that come with AI – existential, political, legal – need time to be worked through properly. I’m skeptical of the alarmist framing that says everyone must drop everything and reorganize their lives around AI immediately.

The mindset that helps a lot in this age of AI is not treating your credentials as your authority. The traditional model of professional expertise worked like this: you learned something at university, consolidated it through years of practice, and your authority grew slowly through experience. AI disrupts that, because the person who tends to win here is often simply the one most willing to engage, to tinker, to keep learning. A twenty-year-old with genuine curiosity and commitment can, at the purely practical level, outperform someone with decades of credentials – simply because they’re willing to put in the exploratory work.

That’s simultaneously good news and bad news. Bad news if you’ve stopped wanting to develop. Genuinely good news if you’ve felt a bit stuck and are looking for a way back in. What the world needs is people who can take AI and apply it creatively within their specific domain.

What has changed, I think, is whose moment this feels like. The 2010s felt like the decade of the extrovert – collaboration, open offices, working out loud. I think we’re entering a different era. This feels more like the moment of the architect-type: someone who sits down, builds systems, tests them, optimizes them. I even wrote in the book – half-joking, not entirely – that this might finally be the age of the introvert.

(Though I’d add: please still go outside occasionally, and use the time you save with your AI agents to have coffee with actual humans.)

AI slop is everywhere now. Coming from a publishing background, what do you think is the future of content?

A lot of what’s out there right now is genuinely repulsive to me. The obviously AI-generated YouTube videos, the weird artifacts, the uncanny valley stuff. But here’s the thing: that’s not even the scary part. The scary part is what comes next, when the technology gets good enough that you truly can’t tell the difference anymore. The problem won’t be that the videos look wrong. The problem will be that they look right.

But to really answer the question, I need to step back a little. I don’t think it’s useful to simply lament the state of modern content as if everything was better before. Even high-value journalistic content has always been shaped by the conditions under which it’s produced and consumed – and those conditions, when it comes to mass media, have always been economically determined. I wish I could say it was purely about enlightenment values – but it was always a negotiation between the ideals journalists cared about and the logic of capital that funded them.

What’s different today is the widespread content agnosticism. TikTok’s algorithm doesn’t care what is in the videos it distributes. It only cares whether it keeps you in the app. And it turns out that a spinach dish rotating slowly in lukewarm water will do that just as well as anything else, if the engagement numbers say so. When you extend that logic – content-agnostic algorithms optimized purely for attention retention – and then add AI into the mix, the darkest scenario is that AI gets used to produce content that’s precisely engineered to keep you scrolling and never let you go.

That’s the bleak version. But I don’t think it’s the only version.

A friend of mine works at a linguistics institute, and he told me recently that for the first time, Google Gemini is making it possible to do systematic high-quality handwriting recognition on historical manuscripts. Entire archives that were previously inaccessible are suddenly opening up. That’s one small example of something I believe more broadly: AI can help us excavate things that already exist – in archives, in overlooked corners of human knowledge – and surface things people simply missed.

The role I find most compelling is AI as a sparring partner. When you work in a niche, it can be hard to find people willing to think something through with you. AI fills that gap in a way that feels useful.

But the real question isn’t what AI is capable of. It’s what goals we’re pursuing. If the structural goal is to addict and dumb down, AI will accelerate that. If the goal is genuine enlightenment – helping people think better – then AI becomes a precision tool in service of something worth doing.

If you were 18 right now, what would you study or do?

I’m in a lucky position today in that I get to combine two sides of myself – the humanities-trained, creative side that’s drawn to human psychology, and the more technical, systems-oriented side. I currently work closely with data scientists and software engineers, and even though AI has become a surprisingly effective bridge between our different ways of thinking, I still sometimes wish I had more of their foundational knowledge. Not to become a software engineer myself – just to not need everything explained to me.

So if I were 18 today? I’d probably still study philosophy – it builds exactly the things that matter most and that AI can’t easily replace: critical thinking, analytical rigour, comfort with ambiguity. But I’d pair it with something like data science or computer science. Not to become deeply technical, but to develop that bilateral fluency earlier. To be able to move between worlds more naturally.

I think that combination – humanistic depth plus enough technical literacy to be dangerous – is one of the more valuable things a person can cultivate right now.

What’s your one piece of advice for someone feeling overwhelmed by AI and not knowing where to start?

For private use, my advice is actually pretty simple. The two biggest mistakes I see people make are either never really getting started – endlessly reading about AI, discussing it in the abstract, meaning to try it someday – or going the other direction and signing up for ten different specialized tools that are essentially just a general-purpose language model with slightly different packaging and a monthly fee.

My actual recommendation: spend the few euros a month on a subscription to a good general model. Claude is my current favourite. And then just start using it. Not by reading more about how to use it – by actually using it. Ask it what you could use it for. Let it show you. Notice where it fits naturally into your day, and keep asking yourself in small moments: could AI help me with this right now?

For organizations, the picture is a bit different. Individual experimentation matters, but what actually drives change is having a clear framework to operate within, where leadership doesn’t just tolerate AI use, but actively expects it. The organizations that make real progress aren’t the ones with the most tools. They’re the ones where trying things with AI isn’t a side hustle you do quietly; it’s part of how work gets done.

You can connect with Franz via LinkedIn, his (German-speaking) website, or by buying his book: “Wie KI dein Leben besser macht”. Thank you Franz for having this conversation with me!