Likeness Notes: An AI-Realized Concept Album as a Field Experiment in Declared Intent, Cybernetic Workflows, and Methodology

Abstract

Likeness Notes Synopsis Likeness is an independent, thirteen-track Indie Folk concept album realized entirely through multimodal generative artificial intelligence (Google Gemini 3 and Lyria 3), accompanied by an exhaustive monograph. Rather than treating artificial intelligence as a push-button novelty or a tool for automated commercial slop, this repository operates within the traditions of practice-as-research and critical making. It investigates what happens when we reject volume and velocity as metrics, replacing them with rigorous aesthetic intentionality, manual audio engineering, and ethical accountability. Likeness offers an unredacted audit of an artistic encounter with an algorithmic black box: disclosing instructions, failed tracks, lyric iterations, mixing decisions, and legal considerations that shaped the final record. UPC: 882436992438 Core Inquiries Disclosing AI Usage The EU Artificial Intelligence Act, Art. 50 (4) focuses strictly on deepfakes, leaving generic synthetic audio in a regulatory vacuum: When institutional frameworks fail to distinguish between automated slop and intentional curation, what does meaningful, proactive disclosure look like in practice? The Threshold of Human Intent Where does algorithmic probability end, and where does genuine artistic labor begin? Can an operator establish authentic authorship when working with stochastic models? Subverting the Median How can an artist steer generative audio engines away from their default statistical mean—the polished, generic clichés of commercial pop—toward idiosyncratic, emotionally resonant sonic territory? Artificiality as Narrative Substance Can the synthetic nature of generative media be reclaimed as an integral dimension of the artistic vision rather than concealed as a technical flaw? A Blueprint for Independent Creators How can solo artists establish sound recording rights (*[Verwertungsrechte des Tonträgerherstellers]((https://www.gesetze-im-internet.de/urhg/__85.html))*), protect their work against automated takedown bots, and implement ethical revenue-sharing models while navigating streaming monopolies? Core Proposals Documented Labor as DemarcationCreators who employ generative models as authentic instruments must consciously distinguish themselves from low-effort slop through proactive, rigorous documentation—an investment of time and craft that anonymous spammers fundamentally cannot replicate. A Native Provenance URL in the DDEX StandardDigital streaming platforms and distributors should introduce a dedicated, standardized metadata field within ingestion schemas (such as DDEX ERN). Providing a direct outbound link to a verified dossier or repository creates an accountable link between an ephemeral synthetic track and its human provenance. Shifting the Discourse from Tooling to Human IntentJust as low-level software syntax is declining in relevance, cultural and critical discourse must move beyond the trivial question of whether AI was employed, focusing instead on how and to what purpose. Evaluating that how, however, requires substantive, open provenance rather than blunt, binary platform badges. Audio & Album Access Likeness is distributed to all major Digital Streaming Platforms (DSPs), released on September 25, 2026. Simply search Jan Köster Likeness Changelog New Document: Release Follow-Up Refined argument on transparency and AI labeling Q and A: comment concerning loudness, "details" for easier access