74 %
of companies see no tangible value from their AI investments.
BCG AI Radar 2025, 1,803 C-level respondents
Research and publications
I test methods on real projects before I teach them.
Empirical work with practical wisdom: findings with their sources, publications with their dates and a framework that has matured in 63 documented versions since 2022.
Starting point
Why I do research at this point
The bottleneck rarely sits in the technology. It sits where values remain unspoken and therefore never become requirements.
10 · 20 · 70
of the value of an AI programme: 10 % algorithms, 20 % data and technology, 70 % people and process.
BCG, the 10-20-70 rule
80 to 8
per cent: leaders who believe they deliver a superior experience, against the customers who agree.
Allen, Hamilton & Reichheld 2005, Bain & Company
70 %
of digital transformations fail. All ten documented causes are human and process causes.
George 2024, Forbes
18.8 to 17.3
per cent hope for career progress through AI, almost as many fear losing their job.
Anthropic Economic Index 2026, 80,508 respondents across 159 countries
84
AI ethics documents have been catalogued worldwide. Almost none of them carry an executable process.
Jobin, Ienca & Vayena 2019, Nature Machine Intelligence
Research position
Contextual constitutions for AI systems
An AI system acquires a character, whether we design it or not. My work gives organisations a procedure with which they determine that character deliberately and can evidence it.
General catalogues of principles help an organisation very little. They state that a system should be fair and leave open what fairness means in this context, for these stakeholders, with this data situation. I therefore work along ISO/IEC/IEEE 24748-7000:2022: values are elicited, weighted and translated into testable requirements that a development team can implement and an audit can read.
The procedure is called the ADA AI Ethics by Design Framework™. It has matured in 63 documented versions since 2022, has been tested on projects in public administration, infrastructure and education, and is available as software through the goodsouls.AI platform.
Current status
From Constitutional AI to Contextual Constitutions
The position paper is available in version 8.0: 36 pages, with an open register schema under CC BY 4.0 and a case analysis of the July 2026 incident based on primary sources. I am completing the submission and will publish the version here. Write to me if you would like to read it in advance.
Add me to the listPublications
What I have published
Journals, case studies and white papers, each with its place and date of publication.
2026
Position paper
From Constitutional AI to Contextual Constitutions
My argument: an AI system needs a constitution that comes from the context of its stakeholders, and a general catalogue of principles does not provide one. The paper is under academic revision.
Under revision
03/2025
Case study
datahub.Tirol — Valuebased Engineering in public administration
Published by IEEE Europe together with Standortagentur Tirol. The project received the Constantinus Award 2025 in the Ethics & CSR category.
11/2024
Expert article
AI — data, ethics and law, illustrated by datahub.Tirol
Published in IT Welt together with Georg Huber.
2024
Expert article
ISO 42001 — the compass for trustworthy AI under the EU AI Act
Published in the journal Comply.
2024
White paper
From Digital Humanism to Value-Based AI Systems
The bridge from digital humanism to a procedure that translates values into testable requirements.
2024
Interview
AI Competence & Digital Literacy in Education
For Österreichischer Bundesverlag.
On request
White papers
I send you these two papers as a PDF.
White paper
Happy Together — Gemeinsam besser
New work, new thinking
How collaboration changes once generative AI becomes part of the team, and which values have to carry that change.
Request the PDFWhite paper
KI im Klassenzimmer
How artificial intelligence changes education
An orientation for schools, teachers and education providers: opportunities, risks and the question of responsibility.
Request the PDFOngoing work
Where I publish
AI Value Bits & Bytes
My weekly briefing on generative AI for decision makers. What has happened, and what it means for your role.
Read the newsletter ↗CuiBono.AI
The podcast on the question of who benefits from an AI system. Conversations about ethics, leadership and technology.
Coming soongoodsouls.AI
The platform where the method becomes software: the ADA sprint, SophiA and the Academy.
Visit the platform ↗Commissioned work
White papers, case studies and reports
You need a robust publication on AI, ethics or governance? I write it with a full set of sources, a visible method and a statement that holds.
What is included
Full set of sources
Every claim with its reference, every figure with year and provenance.
Visible method
How I arrive at the result stands in the document, not in an afterword.
Standards
ISO/IEC/IEEE 24748-7000:2022, ISO/IEC 42001, ISO/IEC 42005 and the EU AI Act, where they carry weight.
Typesetting and delivery
PDF and DOCX for further editing, in your brand layout or in mine.
Let us talk about your project
Tell me your starting point, your goal and your timeframe. I reply with an assessment and a proposal.