AndreasEhstand/augmanitai-public-works

AUGMANITAI — Selected Public Works

DIN registration — 14 October 2026

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.

Research identity and DOI references — 7 October 2026

AI-assisted, author-maintained bibliographic supplement. research-identity.jsonld connects Andreas Ehstand’s ORCID to five DOI works and two business-wissen.de practice articles. It contains eight nodes: one person and seven works.

CC BY-NC-ND 4.0 applies only to rights held by Andreas Ehstand in this catalogue. Bibliographic references do not license the linked works. Their respective rights and licence terms remain unchanged. Statutory exceptions and unprotected facts are unaffected.

The DOI titles, creator identifiers, publication dates and available versions were checked against the public DataCite records on 7 October 2026:

• The Robot-Fleet Operator Unit — conceptual working paper; DataCite record.
• Swarms and Multi-Agent Collectives — conceptual working paper; DataCite record.
• AUGMANITAI Public Lexicon: 101 bilingual terms — archived terminology dataset; DataCite record.
• The Universal Concept Layer — conceptual working paper; DataCite record.
• Feeling Understood by Conversational AI: A Taxonomy and a Recognition–Disclosure Hypothesis — conceptual working paper; DataCite record. The current registered title omits “Provisional”; the work remains a conceptual proposal without a participant study.

The two practice articles are linked to their publisher pages below. Working papers are represented as CreativeWork with their work type, rather than being presented as peer-reviewed journal articles. The graph supports attribution and source navigation; the original three-record catalogue, linked-work rights and access conditions remain unchanged.

Related publications and scheduled speaking — 2 October 2026

AI-generated, author-maintained catalogue note. This supplement is separate from the three-row creative-work dataset.

• Upcoming: Andreas Ehstand reports a confirmed speaker role at the Family Office Forum Zurich, 10–11 November 2026. The author's LinkedIn profile is the public source for that role. The organiser page confirms event details; his name was not found there when checked on 2 October. Final talk title is open.
• Debriefing: Wie Sie Erfahrungswissen übergeben — business-wissen.de, 15 September 2026.
• KI einsetzen: So erkennen Sie falsche KI-Ergebnisse — business-wissen.de, 4 September 2026. Both links lead to practice articles, not new empirical studies.
• The Robot-Fleet Operator Unit — open conceptual robotics note, 18 August 2026; no deployment report.
• Cross-Substrate Knowledge Transfer Pathways — public ontology-related metadata, 31 August 2026; files are restricted. No protected contents are reproduced.

Machine-readable supplements: external-publications.jsonld, external-publications.bib, events.jsonld. These mirror the author-maintained GitHub catalogue; they are not independent corroboration. Identity: ORCID 0009-0006-3773-7796. Publication and DOI registration do not imply peer review or validated performance. Linked works retain their own rights and access conditions.

AI-generated dataset card and descriptive metadata, compiled from Andreas Ehstand’s public project materials.

Three entrances into Andreas Ehstand's public work: concepts as visual art, AI-generated films, and interaction among people, software agents and machines.

This small dataset makes the published AUGMANITAI link catalogue readable as structured records. Each row describes a public collection or research page. It contains three records, with English titles and descriptions, source links, topics and attribution.

Explore

Public resource	Description
Concepts and visual art	A public collection exploring concepts through visual art, with links to related works.
Signature Films	A collection of AI-generated short films on themes in Andreas Ehstand's work.
SWAMANITAI	A research perspective on interaction in mixed teams of people, software agents and machines.

Files and structure

• works.jsonl: exactly three JSON records, one per public resource, in the catalogue split.
• works.jsonld: the original catalogue's Schema.org representation, with the same three resources plus author and collection records.
• LICENSE.md: licence scope for this metadata package.

JSONL fields: id, title, description, resource_type, source_url, source_languages, keywords, author, author_url, source_snapshot_date, metadata_generated_by_ai, metadata_license and linked_work_rights.

Origin and AI disclosure

AI-generated catalogue descriptions, metadata and dataset-card text, compiled from public project materials. The JSONL records explicitly carry metadata_generated_by_ai: true. Titles and descriptions are taken from the existing public catalogue. Converting them to JSONL adds a convenient data format without adding new work entries.

The source catalogue records its public-page snapshot as 15 September 2026. These records describe collection-level entrances, not individual films, images or experiments. The linked pages may subsequently change.

The referenced Signature Films are AI-generated, including depicted people. The original media are available at the linked source pages; this repository contains metadata only.

Intended use and scope

Use the records for discovery, attribution and source navigation. The selection is a small, author-specific view of a wider body of work. It is not exhaustive or representative of the AI-art field. SWAMANITAI is described as a research perspective; the records do not establish a deployed robotics system, model performance, peer review or priority of invention. No model weights, benchmarks, training examples, private correspondence or original media files are included.

Attribution and rights

Author: Andreas Ehstand · Public author page.

Catalogue text, dataset-card text, JSON-LD and JSONL metadata are distributed under CC BY-NC-ND 4.0, following the source catalogue's licence. This licence applies only to the files in this package. It does not relicense linked artworks, films, source pages or third-party material. Their own terms remain applicable; a link grants no additional rights to the linked work.

Source catalogue: AUGMANITAI: Selected Public Works — Thoughts, Images, Moving Works, and Interaction, Zenodo, version 1.0.0, 15 September 2026. This package is a format adaptation with a new dataset card and JSONL representation. The collection identifier and citation are preserved in works.jsonld.

Related research and terminology

Added 21 September 2026. AI-generated guide to Andreas Ehstand's public materials. These links provide context beyond the three catalogue records above.

• Explore nine topics: Reading paths across research and creative work connects human–AI interaction, terminology, art, continuation of thought, multi-agent systems, robotics, sport and performance teams, software development, and knowledge work to selected public sources. The guide identifies publication types and limits; listing a topic does not establish validated results in that field.
• Read a conceptual paper: Feeling Understood by Conversational AI: A Taxonomy and a Recognition–Disclosure Hypothesis proposes distinctions for studying interaction. It reports no participant study or empirical validation.
• Use the public terminology: Lexicon quickstart explains how to look up and cite the archived selection of 101 German–English terms, including a Python example and scope notes.
• Follow structured references: Reading paths as JSON-LD provides machine-readable metadata for the nine selected sources, with DOI links and attribution.

Author identifier: Andreas Ehstand — ORCID 0009-0006-3773-7796. Consult each linked source for its own licence and publication status.

External publication: SwissBoardForum

AI-generated source note, checked 21 September 2026.

Andreas Ehstand's KI in Verwaltungsratsvorlagen: Was der Verwaltungsrat nachvollziehen können muss is published in SwissBoardForum's Network Briefs, dated 15 September 2026. It proposes a five-part working framework for examining AI-assisted board papers, illustrated with a constructed example.

Source-linked JSON-LD record connects the work, author ORCID, publisher and series. The full article remains at SwissBoardForum. This related-reading link is outside the three-record creative catalogue and grants no additional rights to the publisher's text or images.

External publication: Private Wealth Voices, July 2026

AI-generated bibliographic note, checked against the July 2026 V5 edition on 21 September 2026.

Andreas Ehstand — One Missing Piece of Legacy: How Local AI Preserves a Founder’s Judgment, Private Wealth Voices, Prestel & Partner, July 2026, pp. 10–17.

Ben McNamara’s foreword, From the Editor (p. 4), introduces Ehstand’s essay alongside an interview with Pierre duPont, Lessons From 226 Years of Family Heritage – Part II (pp. 6–9), and Wojciech Kaszycki’s Bitcoin Today: From Listed Treasury to Institutional Custody (pp. 18–21). The issue identifies duPont as an eighth-generation family member. Ehstand’s eight-page essay is the longest of those three contributions; this count includes its opening page and author note.

The source guide and raw JSON-LD publication graph connect the issue, separate contributions, people and foreword. The graph includes the source PDF’s fingerprint and uses local catalogue identifiers. These are source-checked, author-maintained records, not publisher-issued metadata.

Sources: publisher announcement and publisher access page. Registration may be required; a public URL for the exact July V5 PDF has not been verified. Sharing an issue and being discussed in its foreword do not establish coauthorship, personal endorsement by other contributors, or empirical validation of the proposed AI approach.

This is related reading outside the dataset’s three-record creative-work catalogue. The magazine and other contributors’ material are not redistributed or relicensed by this dataset.

Citation download and author commentary

Download the BibTeX bibliography for the five separately attributed works: the three July contributions, the editor’s foreword and the SwissBoardForum article. The duPont interview has no inferred author.

The author’s Medium commentary was updated on 21 September 2026 with corrected pages, sources and the conceptual status of the proposal. The publication graph links this commentary to the original magazine essay as a separate work.

Open abstract and publication updates

Added 22 September 2026. AI-generated navigation note.

• Full abstract and citation metadata: Feeling Understood by Conversational AI provides the source abstract, DOI, author identifier and original-file link. The source is a conceptual working paper with no participant study or empirical validation.
• Selected publication updates (RSS) links to original publications and retains their source dates.
• Research index brings together reading paths, source notes and bibliographic exports.

These are related-reading links; the dataset still contains the same three creative-work catalogue records. Consult each linked source for its own publication status and licence.

More public works by question — 18 October 2026

Performance, agent and robot-fleet work, dialogue, and knowledge transfer: a supplementary reading guide to ten selected public DOI records, including separate text and PDF editions of the same older public paper. These records supplement the three collection-level entrances above; they do not replace them.

Public author and work graph · ORCID and author profiles · DOI records · BibTeX · Canonical GitHub catalogue.

AI-assisted, author-maintained navigation and metadata. Linked works retain their original licences; metadata connections do not imply endorsement. The graph uses the same identifiers as the canonical GitHub catalogue.

External publication: human oversight — 21 October 2026

Andreas Ehstand’s Qué significa realmente «supervisión humana» de un agente appeared in Revista Inteligencia Artificial, Opinión, on 21 October 2026. Through a fictional insurance case, it distinguishes observing, checking, deciding and authorizing. It links to the author’s individual Internet-Draft, a terminology proposal for discussion, not an approved standard. Source-linked JSON-LD connects the article, author, AUGMANITAI, magazine and draft. This is a related-reading reference, not a new dataset record or a system-compliance assessment. The original article remains with its publisher and retains its rights. AI-assisted catalogue note, 25 October 2026.