Skip to content
The tech theologian's podcast · since 2021
In Ottes OhrDer Podcast vom Techtheologen
Medical assistant on the phone at the front desk
Guide

The human measure: On AI in the doctor’s practice, resistance and our oldest ethics…

May an AI answer the practice phone? Who owns our health data? And why is good technology not enough on its own? In my guest lecture at HTWG Konstanz I tell what happened when I digitised my wife’s gynaecology practice in a short time. Nine theses came out of it, summarised here in plain language.

By Jan Thomas Otte5 min read · 60 min listen
Start the story

Listen to this episode

60 minutes · Audio in German · Guest lecture at HTWG Konstanz, 17 November 2026, lecture series Applied Ethics

Also on: Spotify · Apple Podcasts · podcast.de

May an AI answer the practice phone? In the lecture series “Applied Ethics” at HTWG Konstanz on 17 November 2026 I told what happened over the last two years when I made my wife’s gynaecology practice fit for digital. No product pitch and no fuck-up night, but an honest field report. Instead of slides there was an ethics board on Miro: the students voted live on my theses.

Jan Thomas Otte in a workshop with the practice team at the flipchart

From the practice: workshop with the team – “What is not working well yet?”

Title: Where “AI comes before K.O.” comes from

The German pun “KI kommt vor K.O.” comes from a panel at cyberLAGO at Zeppelin University. It has stayed with me ever since. In the practice, AI came first, then the knockout – and then a second, better attempt.

Theses 1 and 2: The problem runs deeper than technology

My practice was never analogue. Computers had been there for a long time; they were just poorly connected. Hence my second thesis: digitalisation is the consequence of the problem, not its cause. The German healthcare system is considered expensive and not very efficient, and new software alone changes little. Compare Switzerland: insured people first pay part of their medical costs themselves every year. In Germany, a doctor’s visit is free for people with statutory insurance, apart from their contributions.

Thesis 3: A practice is not a delivery service

When a doctor’s visit costs nothing, some expect a practice to work like a delivery service: book online, get served at once. Yet people with statutory insurance have a contract with their health insurer, not with the doctor. So a practice may also say: we are full right now.

“Whoever introduces AI to the max is not testing the technology but the people.”

Theses 4 and 5: AI tests the people

We introduced many programs at once: an online calendar, digital forms instead of clipboards, speech recognition, text reminders and an AI phone assistant. Each made sense on its own. But they did not work together, and the old paper processes kept running. Before, callers sometimes waited up to 45 minutes on hold; afterwards came complaints and bad reviews. My lesson: whoever introduces AI on a large scale is not testing the technology but the people. Whether patients and team come along shows right at the reception desk. Resistance is not a diagnosis in the sense of a disease. It is feedback, and whoever takes it seriously can learn from it.

Sailing: From the yacht into the dinghy

I have led large IT projects in corporations. Still, things went differently in the small practice. I compare it to sailing, which I learned at HTWG, by the way: a yacht lies calmly in the water. In a dinghy, even a small shift of weight can capsize the boat. That is exactly what happened to us in the second year.

“For me, an IT transformation in a doctor’s practice was like stepping from a yacht into a dinghy.”

Theses 6 and 7: The human measure is in no specification

IT projects have a specification, a list of what the software should do. What is reasonable for people is usually not in it. That is why you rarely find the human measure in the first release. Plan the second one right away. We have rolled some things back, and patients can give feedback via a link. The seventh thesis – “good AI goes unnoticed” – I put up for debate on purpose. Since August 2026 the European AI Act requires that people learn when they are talking to an AI, especially with a synthetic voice.

My concrete advice for AI on the phone: two attempts with the machine. On the third attempt, a human should be reachable within a minute, during opening hours.

Thesis 8: Whoever manages appointments holds power

Large booking platforms for doctor’s appointments solved a real problem: getting an appointment is easier. But findings and data of whole families accumulate there, and some platforms have grown so big that they hardly open interfaces to others. My thesis: whoever manages the appointments ends up managing not the appointment but the patient. On the other side is the state electronic patient record (ePA), created automatically for people with statutory insurance unless they object.

Thesis 9: Responsibility stays with people

A machine has no conscience and no intuition. That is why responsibility cannot be handed to an algorithm. Here I draw on the computer scientist Katharina Zweig and her book “Ein Algorithmus hat kein Taktgefühl” (An algorithm has no tact). My concern is less AI itself than what it does to us – for instance when we agree to everything out of convenience. I call that digital overreach.

“We are not writing a new ethics. We are presenting a new technology with our oldest ethics.”

Ethics: Our oldest ethics for a new technology

The pharmaceutical and technology company Merck has written down five principles for dealing with data and AI: justice, autonomy, beneficence, non-maleficence and transparency. Four of them come from the medical ethics of the 1970s and are older than any computer. As a theologian I trace them back further, through the Ten Commandments to the Code of Hammurabi. Only transparency is new. It is uncomfortable because it raises the question of power: how far am I willing to show my cards?

Food for thought: What is the human measure?

I have no finished answer. I hold on to a definition from the book “Digitale Ethik” by Grimm, Keber and Zöllner: digital ethics looks for appropriate rules and values for how we act in a digital world. My wish is clear: people should remain reachable, even when a system takes part in the decision.

Sources and further reading

Photos: Alexander Stertzik (workshop); Getty Images via Unsplash+; Daniel Stenholm / Unsplash