Short answer: she doesn’t exist. And that’s the point.
Sarah is the host of this podcast. The show carries her name. Her voice is AI-generated. Her personality is a writing device. Every question she asks, every objection she raises, every summary she delivers was researched, written, and signed off by a very real human: me, Markus.
This page explains why the show works this way, how it’s made, and where the boundaries are.
Why a synthetic host?
Because good arguments need resistance. I’ve worked in the internet business since 2000, in classical media before that, and if those years taught me one thing, it’s this: a monologue is where thinking goes to relax. An expert talking into a microphone, unchallenged, gets comfortable. Comfortable is boring — and worse, comfortable is often wrong.
Sarah is my built-in contradiction. She asks the questions a smart, skeptical listener would ask. She brings up the numbers that don’t fit my thesis. She sets traps I walk into on air. When I claim AI makes me twenty percent faster, she’s the one holding the stopwatch.
Could a human co-host do that? Absolutely — and better in many ways. But this format lets one person produce a genuine dialogue: I write both sides, and writing the strongest possible case against my own position has made every episode sharper. Sarah exists so that nothing on this show goes unchallenged. Not even Sarah.
How it’s made
Each episode starts as research, then becomes a written dialogue script. I write all of it — Markus’s lines and Sarah’s. Her voice is then produced with text-to-speech technology, with stage directions for pacing and tone. Nothing Sarah says is generated live or autonomously. There is no AI improvising opinions on this show. If Sarah is wrong about something, I was wrong first.
Editorial responsibility for every word — hers and mine — lies with me. That’s not a formality. It’s the standard that separates a produced show with a synthetic voice from an automated content feed.
Why tell you all this?
Two reasons.
First, you deserve to know who you’re spending time with. Listeners build a relationship with voices. That relationship should stand on honest ground. A synthetic voice isn’t a deception — an undisclosed synthetic voice is.
Second, we cover these rules — so we follow them first. Since August 2, 2026, Article 50 of the EU AI Act requires transparency about AI-generated content, including synthetic voices. This show reports on that regulation regularly. It would be absurd to explain labeling duties to you through an unlabeled AI voice. So Sarah discloses herself at the top of the show, every episode. We were doing it before it was law. Now it’s both.
What Sarah is not
To keep things honest, the boundaries in plain terms:
Sarah is not a real person, and no real person served as a template for her. She is not an autonomous AI with her own views — she has exactly the views I write for her, including the ones I disagree with. She doesn’t answer messages, give interviews, or appear at conferences (though the invitation would flatter both of us). And she is not an attempt to replace human co-hosts, journalists, or anyone else. She’s a format choice by one producer — disclosed, scripted, and accountable.
The name
Sarah has a last name. I have a soft spot for a specific country and its music and vibes, and the name sounded like someone who would interrupt me mid-sentence with a better argument. That felt right.
Questions, criticism, curiosity
If you have thoughts about this format — whether it delights you or bothers you — I genuinely want to hear them. Synthetic voices in journalism-adjacent formats are new territory, and the norms are being written right now, partly by shows like this one and listeners like you. Write to me via the contact page.
Sarahs Tech — a show hosted by someone who doesn’t exist, with facts that very much do.
