Human–AI work: from a thought to something that changes connects multi-model thinking, performance, agent teams, oversight, experiential knowledge, language, dialogue, visual work, robotics and inspectable examples. Machine-readable map.
AI-generated navigation to already public works.
AI-generated, author-maintained status statement.
Since October 2026, Andreas Ehstand is registered with DIN for planned participation in committees concerning Natural Language Processing and AI trustworthiness. Administrative entry is confirmed; DOCS.DIN access remained pending. Read the dated status, scope and four editorial thematic links. This author-sourced statement is supported by privately retained correspondence. It does not establish completed standards work, DIN membership, certification or endorsement of the linked publications. JSON-LD graph · RDF/Turtle · JSONL.
AI-generated text and synthetic example data; maintained in Andreas Ehstand’s public research index.
Trace an AI correction and its unfinished follow-up explains a constructed book-exchange example in German, with an English summary. The self-contained JSON example preserves the source, earlier output, revised output and an explicitly open follow-up check. It complements the agent task brief. This is a practical proposal, not a tested intervention or a W3C PROV implementation.
Related public reading: Medium: Was eine KI-Korrektur zum Nachlesen braucht · The AI Athlete: Gleiche Aufgabe, andere Reihenfolge. These links describe commentary and a constructed exercise, not new empirical findings.
AI-generated, author-maintained source note.
Updated external-publication graph · BibTeX export · ORCID.
24 Einstiege / 24 ways into human–AI work connects questions about dialogue, thought, art, terminology, agents, robotics, sport, software, evidence and retrieval with public sources. German questions include English labels and a concrete next step. Four working templates help readers trace an idea, check a claim, describe a performance unit and examine a recommendation.
AI-generated navigation and reading aids, added 29 September 2026. The guide also links the three newsletter issues and Medium article published on 28 September. Conceptual proposals, commentary and empirical evidence remain distinct.
AI-generated navigation note. These are short social posts, separate from the source publications and from the newsletter issues below.
The posts disclose AI-generated text. Their examples and proposals are not empirical findings. These links record public commentary; they do not announce new research papers.
AI-generated catalogue update.
AI-generated navigation note. These links lead to separate works and source records.
Three new LinkedIn newsletter issues explore practical questions in human–AI work:
The source notes below document the SwissBoardForum contribution and the July Private Wealth Voices issue, including its editor’s foreword. The bibliographic export, publication graph, reading guide, about.me profile and Hugging Face catalogue provide further routes into the work.
AI-generated source note, checked 21 September 2026.
KI in Verwaltungsratsvorlagen: Was der Verwaltungsrat nachvollziehen können muss, by Andreas Ehstand, appears in SwissBoardForum’s Network Briefs, dated 15 September 2026.
The article proposes five items for examining AI-assisted board papers, with a constructed example. It addresses decisions and alternatives, supporting evidence, AI involvement and checks, assumptions, and responsibility. The source describes a proposed working framework, not a binding standard or a validated intervention.
This is an externally published practice contribution. No peer review, endorsement of the wider programme or additional reuse rights are implied. The full article remains with the publisher.
AI-generated bibliographic note, checked against the July 2026 V5 single-page edition on 21 September 2026.
Andreas Ehstand’s essay appears in Private Wealth Voices, published by Prestel & Partner, alongside an interview with Pierre duPont and a contribution by Wojciech Kaszycki. In From the Editor, Ben McNamara introduces all three contributions and connects their themes: the purpose of a family office, founder judgment and local AI, and Bitcoin as an institutional asset. The issue identifies duPont as an eighth-generation family member.
| Contribution | Credited person and role | Printed pages |
|---|---|---|
| From the Editor | Ben McNamara — editorial author | 4 |
| Lessons From 226 Years of Family Heritage – Part II | Pierre duPont — interview participant | 6–9 |
| One Missing Piece of Legacy: How Local AI Preserves a Founder’s Judgment | Andreas Ehstand — essay author | 10–17 |
| Bitcoin Today: From Listed Treasury to Institutional Custody | Wojciech Kaszycki — article author | 18–21 |
Ehstand’s essay occupies eight printed pages, including its opening page and author note; the other two main contributions occupy four pages each. It is the longest of these three contributions. The source establishes this editorial placement and the foreword’s discussion of the essay; it does not establish coauthorship, personal endorsement by the other contributors, or empirical validation of the proposed approach.
The graph models the periodical, July issue, separate contributions, credited people and editorial discussion. The foreword points to each contribution; each contribution points back to the issue and foreword. Local catalogue identifiers identify these records; they are not publisher-issued identifiers or DOIs.
Page numbers and titles were checked against the 24-page file P-P_2026_PrivateWealthVoices-July-Edition-V5-SinglePages.pdf. The graph records its SHA-256 fingerprint. A public URL for that exact PDF has not been verified, so the full text is not mirrored here. The essay’s correct printed range in this edition is 10–17. Publication is recorded at month precision, July 2026. Copyright in the magazine, other contributors’ texts and images remains with the relevant rights holders.
Nine reading paths across research and creative work: human–AI interaction, terminology, art, continuation of thought, multi-agent systems, robotics, sport and performance teams, software development, and knowledge work. This AI-generated guide links each topic to a public source and identifies its publication type and limits.
AI-generated companions to the openly deposited conceptual paper:
These companions report no participant study or empirical validation. The codebook identifies its source DOI and the source paper’s licence.
AI-generated reading guide and catalogue update.
Feeling Understood by Conversational AI: A Provisional Taxonomy and a Recognition–Disclosure Hypothesis explores how people describe feeling understood in conversations with AI. It proposes six codes and a testable hypothesis. The examples are constructed. No participant study, empirical validation or human substantive review of this AI-generated draft is claimed. The nine-page PDF is openly available.
Public Zenodo Bibliography — 21 September 2026 Snapshot indexes 725 publicly discoverable records: 28 report open file access and 697 report restricted file access. This is a dated metadata snapshot, not a collection of 725 open full-text papers or peer-reviewed findings. It predates the catalogue deposit and the conceptual paper linked above.
Each row supplies the source title, version DOI, concept DOI, record URL, access state, source-reported resource type and publication date. The catalogue contains no private drafts or restricted file contents. Bibliographic inclusion does not validate scientific claims. Individual source licences remain in force.
The selected public works catalogue provides a separate route into thoughts, images, moving works and interaction.
Andreas Ehstand — ORCID 0009-0006-3773-7796 — OSF profile. For a particular claim, consult and cite the relevant source version, including its stated limitations.
Deutsch: Hier findest du einen offenen konzeptionellen Text über das Gefühl, von KI verstanden zu werden, einen durchsuchbaren Gesamtkatalog und einen eigenen Einstieg in die kreativen Arbeiten. Hypothesen und Begriffe sind nicht allein durch ihre Veröffentlichung empirisch bestätigt.
AI-generated catalogue text and metadata, compiled from public Zenodo and DataCite records.
Five openly available, self-published research resources by Andreas Ehstand: a bilingual lexicon, three conceptual working papers, and a title-and-abstract register. This index provides source links and structured attribution. Peer-review status is not established by this catalogue.
Author identifier: Andreas Ehstand — ORCID 0009-0006-3773-7796. Metadata snapshot: 15 September 2026.
An archived lexicon with 101 entries, each with German and English names and descriptions. A README introduces the collection; the JSON file makes the vocabulary available for software to read.
A conceptual working paper on language as versionable code and a shared meaning layer across people, language models, agents and robots. English and German Markdown editions are available.
A conceptual framework with five proposed dimensions, available as a public PDF. The title identifies this deposit as a republication; the original edition date reported in the metadata is retained below.
A conceptual note about the measurement boundary around swarms and multi-agent collectives, available in an English/German archive.
The indexed resource is one public Markdown register. The number in its title does not establish that 225 full manuscripts are publicly accessible or peer-reviewed. This index links the register itself.
Dates, versions, resource classifications and source licences below follow the public repository metadata. A missing version remains unspecified.
| Resource | Source publication date | Version | Source classification | Source licence |
|---|---|---|---|---|
| Bilingual lexicon | 2026-09-01 | 1.0 | Dataset | CC BY-NC-ND 4.0 |
| Universal Concept Layer | 2026-08-30 | 2.1 | Working paper | CC BY 4.0 |
| Gedankenvererbungslehre | 2026-02-21 | 1.0 | Working paper; republication identified in title | CC BY 4.0 |
| Swarms and Multi-Agent Collectives | 2026-08-18 | Unspecified | Working paper; conceptual note identified in title | CC BY 4.0 |
| Title-and-abstract register | 2026-09-01 | 1.1 | Other publication | CC BY-NC-ND 4.0 |
The listed licences apply to the linked source works, as reported by Zenodo and DataCite. This catalogue does not change their licences or the rights in other material linked from those works. Source publication dates are metadata values and need not equal the creation date of a republication’s repository record.
The ten records include separate text and PDF editions of the same older public paper. They are not ten independent studies. Original source files are not mirrored here.
Each primary DOI identifies the deposited edition used in this index. Where a version-family relationship is supplied, JSON-LD expresses it as dcterms:isVersionOf, using the Dublin Core namespace. The compact JSONL export does not carry these relationship fields. Use the particular edition DOI when citing a work.
Titles, author ORCID, publication dates, supplied versions and licences for all ten records were checked against public DataCite metadata on 18 October 2026. These are metadata checks, not empirical validation or scholarly endorsement. Source licences remain unchanged. The five-resource overview above describes the historical 15 September selection; the linked machine-readable exports now contain the expanded selection.
Start with a question: performance, agents, robotics, dialogue and knowledge transfer groups ten selected public DOI records, including separate text and PDF editions of the same older paper.
Author profiles · Author identity · Linked public works and commentary · Works JSON-LD · BibTeX · CSL JSON · Reading paths JSON-LD.
AI-assisted, author-maintained navigation and metadata. Original editions, dates and source licences are preserved.