Andreas Ehstand · AUGMANITAI
An idea becomes a conversation. The conversation becomes a picture, a piece of software, a decision or a shared task. Something happens outside the conversation, and that response changes the next step.
AUGMANITAI’s ambition is to make this whole process easier to understand, develop and carry across people, models and tools. These public works connect performance coaching, human–AI dialogue, creative work, experiential knowledge, multi-model thinking, agents and robotics.
This guide offers several ways in. Each leads to an original publication or an existing public collection.
What changes when someone develops a sustained way of thinking with several AI models, while keeping their own judgment and testing ideas against consequences outside the chat?
Read the conceptual definition of AUGMANITAI.
An impressive answer is one moment in a longer process. The human still frames the task, notices mistakes, chooses what matters and decides what to do next. How can those contributions become visible?
Four practical questions for analysing your own performance when working with AI transfers a performance-coaching perspective to demanding human–AI work.
Several agents, tools and human decisions can contribute to the same outcome. Where should we draw the boundary around the team before describing its performance?
Swarms and Multi-Agent Collectives explores that question as a conceptual note.
Seeing an output, checking a property, choosing an action and authorising it are different acts. Recording that distinction can make the question of human oversight more precise.
Terminology for Human Oversight Acts in Automated and Agentic Systems is an individual Internet-Draft offered for discussion, rather than an adopted IETF standard.
What needs to travel with a decision so that someone else can understand the thinking behind it?
Gedankenvererbungslehre develops a conceptual framework for the continuation of thought. The practical article Debriefing: Wie Sie Erfahrungswissen übergeben addresses the handover of experiential knowledge.
People and software may use the same word while relying on different definitions. A later revision adds another question: which version shaped this text or decision?
Explore the bilingual public lexicon and the article When a definition changes: which version did the AI-assisted text use?.
Feeling understood is an experience worth describing carefully. How could observations of a conversation and a person’s own account be distinguished and studied?
Feeling Understood by Conversational AI proposes a taxonomy and a research hypothesis.
An image or interactive work gives an idea another form that others can encounter and interpret. These existing collections provide a route into that creative side of the work.
Browse the public works catalogue · Archived catalogue of moving and interactive works.
When robots enter a working arrangement, the world can answer through physical contact: a gripper meets an object or empty space. Which parts of the human–AI–robot arrangement belong in its description?
The Robot-Fleet Operator Unit develops this as a conceptual transfer to robotics.
Meaning and Validity of Stored Memory Items, v0.3 provides synthetic test material with a reference checker. The title-and-abstract register offers further entry points across human–AI interaction; it is a register rather than a collection of 225 accessible full manuscripts.
Start with the question closest to your work. Read its original source, inspect any linked material and identify the part you would test, change or use. Twenty-four further entry points and the existing collaboration guide connect the wider collection.
Scope: These resources include conceptual proposals, practitioner articles, creative collections and synthetic examples. The guide reports no comparative performance result, deployed robot fleet or institutional endorsement. Original sources retain their own dates, editions and licences.
Author: ORCID 0009-0006-3773-7796 · Public research index · Structured reading map
AI-generated navigation text, compiled from already public sources, 18 October 2026.
Andreas Ehstand is a member of the AI Future Council at the Swiss Future Institute. His focus includes Human–AI–Robotics Collaboration, knowledge and experience continuity, and AI-supported holistic performance analysis. Katrin J. Yuan chairs the Council.
Membership announcement · Author identity: ORCID · Structured membership context.
The membership provides author context; it does not constitute institutional endorsement of the works linked in this guide. AI-generated context text, 19 October 2026.