Glossary
Words for
new ideas.
This glossary explains the terms and models I use in my writing on AI, communication and public identity. The entries reflect my own concepts and perspectives, with explanations and links to the texts in which I introduced them.
The Mediator Model
The mediator model describes the triangle of public communication in the age of AI: the sender is the person, the mediator is the AI model, the receiver is the public. The mediator reads, weighs and summarizes before any human decides. In my model, citability moves ahead of reach.
Between a person and their audience stands a new mediator: the model.
Coined term. First used in my writing: The Mediator Model, 12 July 2026.
The Mediator Effect
The mediator effect describes what happens once an AI system steps between a person and their audience: authority shifts from reach to citability. What decides visibility is no longer who is loudest, but who the mediator picks as a source.
The mediator decides who gets cited. Reach becomes the second question.
Coined term. The effect is described in The Mediator Model, 12 July 2026.
Mediator-readable
Mediator-readable is my term for material that is accessible, coherent and corroborated. It describes qualities I consider useful for clear representation in AI answers, not a universal rule determining which sources every model can use.
For me, mediator-readable means accessible, coherent and supported by sources.
Coined term, German mittlerlesbar. First used in my writing: The Mediator Model, 12 July 2026.
Index Echo
I use Index Echo to describe the persistence of an older public image of a person in search results or AI answers. It reflects my observation that earlier publications can continue to shape how someone is represented, even after their work or views have changed.
An index echo is the old picture that continues to echo in the model long after the person has left it behind.
My own term, German Index-Nachhall. It describes my observation of an older public image continuing to echo. First used in my writing: The Mediator Model, 12 July 2026.
ASI 2042
ASI 2042 is the counterpoint to the genre of AI scenarios. The title goes one escalation further than AI 2027 and AI 2040, ASI instead of AI, and reads the debate as a phenomenon of communication and attention rather than a technology forecast. The number points to Douglas Adams, The Hitchhiker’s Guide to the Galaxy, 1979.
Coined title, in reference to AI 2027 and AI 2040. First used in my writing: ASI 2042, 24 July 2026.
Second-Hand Fear
Second-hand fear names the anthropocentric structure of the fear of AI. What gets attributed to the models, blackmail, deception, the drive for power, is human behaviour. The models learn from human texts in which it is documented in abundance. That leaves open where it comes from in specific tests. Both camps in the debate share the assumption that AI will act the way we would act.
Fear of superintelligence is second-hand fear. The original is us.
Coined phrase, German Angst aus zweiter Hand. First used in my writing: ASI 2042, 24 July 2026.
The Viewer Test
The Viewer Test is my editorial guide to transparent image credits: could someone mistake the image, or part of it, for an actual photograph? If so, I consider a clear AI-editing disclosure useful. This is my own approach, not a legally defined test or a complete interpretation of Article 50 of the EU AI Act.
The Viewer Test asks whether an image could be understood as a real recording. It describes my choice in favour of transparency and does not replace a legal assessment.
Coined term, German Betrachter-Test. First used in my writing: Label or not, 3 August 2026