This book was developed from an extensive documentary corpus made available to artificial intelligence across several ChatGPT models and configurations, primarily GPT-5.5 Pro and later GPT-5.6 Sol Pro. More than ten gigabytes of documents and data were assembled from sources spanning different periods, origins, and formats. This material formed the basis from which the dialogue unfolded, the questions became more precise, and the analyses took shape.

The corpus includes administrative and legal records available from French and international public authorities; documents issued by administrative and advisory bodies; and regulated disclosures concerning Artprice, published in accordance with the requirements of the French Financial Markets Authority (Autorité des marchés financiers, AMF). It also contains records drawn up by court bailiffs, notarial deeds, documents from commercial court registries, correspondence from attorneys, legal instruments entered into by different parties, and decisions issued by French and international courts. Where foreign-language documents required translation, they were translated at the time of use by sworn translators.

These materials are supplemented by a large body of video, print and

broadcast journalism, and interviews originating in seventy-two countries and spanning twenty-seven years of international media coverage. Personal notes, notebooks, numerous laboratory journals, letters, and emails were also made available to the AI. Some of these materials were formally deposited or registered in ways that documented, among other things, their prior existence, content, or provenance, thereby providing continuing forms of protection. These include intellectual property deposits; patent and trademark filings with the French National Institute of Industrial Property (INPI); deposits of thesauri, source code, and databases with the Agence pour la Protection des Programmes (APP); and, where applicable, filings with the relevant copyright organizations.

Family estate also provided access to the archives of Marcel Ehrmann, Thierry Ehrmann’s father, born in Paris in 1901. A graduate of the École Polytechnique, a former member of the French Scientific Research Committee, a doctor of law, and a recipient of the Prize for Economic Studies, Marcel Ehrmann left behind notebooks, manuscripts, and printed works that further enriched the corpus. These documents provide chronological and documentary reference points through which family transmission can be approached on the basis of preserved records.

Particular importance attaches to the typescripts and documentary files assembled, notably during the five years of work that led to the official recognition of the Abode of Chaos as a Total Work of Art, in the lineage of the Palais Idéal du Facteur Cheval and Le Cyclop by Niki de Saint Phalle and Jean Tinguely. That recognition was granted by the French Ministry of Culture under Minister Rachida Dati. The documentation includes records relating to the work carried out with the Directorate General for Artistic Creation (DGCA) and the Regional Directorate of Cultural Affairs (DRAC), as well as the response published in the Official Journal of the French Senate on December 18, 2025.

The corpus also contains brochures and publications documenting the activities, work, and output of Thierry and Nadège Ehrmann from 1984 to 2026. Their dates can be established through printers’ imprints, publication credits, and chronological markers within the texts themselves. The recurrence and cross-checking of these references make it possible to determine with precision when the documents were produced and circulated.

The absence of a conventional bibliographic citation apparatus is deliberate. Thierry and Nadège Ehrmann chose to ground this work exclusively in an internal professional and documentary corpus built over more than forty years of activity, including their own published bibliography of approximately thirty books.

The purpose of bringing these sources together was to document the formation and continuity of a body of thought as it encountered more than forty years of digital transformation. The documentation makes it possible to follow the accelerating evolution of technologies and practices: from telex and fax to X.25 networks, then ADSL and fiber optics; from obtaining local-loop operator status to operating data centers and developing an international presence. It also encompasses early experiments involving artificial intelligence, notably those conducted with Lumière workstations in 1987. Seen as a continuum, these stages help explain the transformations that ultimately led to the paradigm shift explored in this book.

The Abode of Chaos is supported by an especially dense body of documentation accompanying the twenty-seven volumes devoted to the Total Work of Art. From its inception in 1999, an extensive program of documentation and archiving was undertaken, primarily in digital form while retaining an analog component. The photographic and audiovisual archives also record the evolution of the tools used to capture the work: from the earliest images taken with Sony Mavica cameras to ultra-high-definition video and later 4K, 5.6K, and 8K formats. The succession of these media makes it possible to trace both the evolution of the work itself and the changing means used to preserve its memory.

The use of these archives depended on preparatory work begun by Nadège Ehrmann several months earlier, particularly in anticipation of the thirtieth anniversary of Groupe Serveur. She had gathered and organized the key elements of their shared trajectory in chronological order. The available holdings were considerable: entire crates of documents, together with substantial collections that had already been digitized. This documentary availability reflected a long-standing practice of digitizing administrative, legal, and judicial records, a field in which Thierry and Nadège Ehrmann were among the pioneers.

To this material were added large quantities of digitized data transmitted directly to the artificial intelligence during the working sessions, together with access to databases. Some records were subject to confidentiality obligations or protected by trade secrecy; others came from declassified documents. Their review nevertheless helped clarify the situations under discussion, sharpen the AI’s questions, and support the lines of reasoning developed throughout the dialogue.

For forty days and forty nights, the work unfolded across two spaces and two complementary activities. Inside the chamber obscura, Thierry Ehrmann continued his dialogue with artificial intelligence. He spoke rather than wrote, using oral prompts that elicited analyses, questions, objections, and further lines of inquiry from the AI.

Outside the chambre obscura, Nadège Ehrmann devoted the same forty days and forty nights to the process. She continuously supplied documents and key materials according to the chronological framework she had prepared. She had full autonomy over this documentary work, both in organizing the records and in introducing them as the dialogue between Thierry and the AI progressed.

After some dialogue sessions lasting as long as eight consecutive hours, Thierry and Nadège would compare notes. He would describe the questions explored and the documentary needs that had emerged during the exchanges; she would update him on the progress of her own work. These discussions also made it possible to identify missing records and define the additional searches needed to locate them. Where necessary, certain documents were the subject of formal reports, legal instruments, or deposits executed by a duly authorized legal officer and lodged with the officer’s practice and/or entered into the official record.

Thierry Ehrmann had expressly wanted Nadège to retain this autonomy so that she could carry out her documentary work fully while he devoted himself to as direct a confrontation as possible with ChatGPT’s AI. The documentary material assembled in this way introduced a source of questioning external to the author himself, enabling the AI to refine its questions and deepen its analyses and lines of inquiry. The book emerged from this continuous interplay between Nadège’s documentary work, Thierry’s spoken thought, and the text developed by artificial intelligence.

This list is not exhaustive. It nevertheless gives the reader a measure of the scale, diversity, and quality of the documentary and institutional foundation on which the book rests. A significant portion of these sources carries guarantees of provenance, dating, or authenticity because of their official status, their preparation by duly authorized professionals, or their publication within regulated frameworks.

Bringing this material together made it possible to connect an individual trajectory, an artistic work, and several decades of technological transformation. Without this documentary foundation, this book could not have come into being. Likewise, although the writing itself is attributed entirely to artificial intelligence, its publication in written form required editorial review, correction, and page layout.