Breakout-Likelihood Rubric for AI Artists and Shows

## Framing **Purpose:** A structured judgment aid for ranking emerging AI and digital artists and exhibitions by breakout likelihood. Not a validated predictive model. **Scope:** Two linked applications: (a) artist breakout, (b) exhibition contribution to acceptance of AI tools in contemporary art. **Theoretical grounding:** - Rogers' *Diffusion of Innovations*: adopter categories, opinion leaders, perceived attributes (relative advantage, compatibility, complexity, trialability, observability). - Becker's *Art Worlds*: networks of cooperating actors, conventions defining legitimate art, distribution and legitimation channels. **Status:** Revised following a retrospective test on Refik Anadol. Revisions to C3 and C5 are required before the rubric is applied to emerging artists (see Revision Log). ## Criteria ### C1 — Opinion-leader network position - **Theory:** Rogers (opinion leaders/change agents); Becker (networks of cooperating actors) - **Measures:** Ties to established curators, critics, institutional gatekeepers who have legitimized work in adjacent media. - **Assessment:** Count distinct gatekeeper nodes (curators, critics, galleries). Score by breadth of distinct nodes, not total mentions. Check reciprocity. - **Type:** Evidence-based - **Speculative note:** Network centrality predicting career outcomes is documented. Whether it predicts AI-specific breakout is untested. ### C2 — Trialability - **Theory:** Rogers (trialability reduces adoption barriers) - **Measures:** Whether non-specialist viewers can encounter the work at low cost (open-access exhibitions, public installations, editions, online interactive versions). - **Assessment:** Inventory accessible encounter points over past 24 months. Score by number and accessibility. - **Type:** Evidence-based - **Speculative note:** Rogers' findings are robust across domains. Mapping to art reception is extrapolation. ### C3 — Convention compatibility (revised) - **Theory:** Becker (conventions define legitimate forms); Rogers (compatibility with existing values/practices) - **Measures:** How well the work fits existing categories (painting, photography, sculpture, installation, performance) versus requiring a new framework. - **Assessment:** Two independent raters place work on fit scale against dominant conventions of target venue. Record disagreements; resolve by discussion, not averaging. - **Scoring note (productive friction):** Low convention fit is not automatically a weakness. Raters must record which of the following applies: - **Weak fit (deficit):** The work does not fit because it is underdeveloped, derivative, or unclear in its medium logic. - **Productive friction (expansion):** The work does not fit because it deliberately extends or hybridises existing categories (e.g., immersive media, architectural projection, data-driven environments) and its own internal logic is coherent. - **Sub-criterion C3b — Productive friction:** Where friction is recorded, rate whether the work shows a coherent medium logic and whether gatekeepers in adjacent fields (architecture, immersive media, performance, design) have engaged with it. Friction with a coherent logic and adjacent-field engagement should not be scored as a convention deficit. - **Type:** Judgment - **Speculative note:** Raters may read compatibility differently depending on medium familiarity. Inter-rater agreement should be reported. The deficit/expansion distinction is a judgment heuristic not yet validated; it was motivated by a single retrospective case. ### C4 — Conceptual and technical legibility - **Theory:** Rogers (complexity as adoption barrier, inverse) - **Measures:** Whether an informed non-specialist can state what the work does and why it matters after a single encounter, without needing technical details. - **Assessment:** Blind read-out: rater reads only artist statement and wall text, then describes work. Score by alignment with artist's stated intent. Invert only where complexity blocks access, not where it adds depth. - **Type:** Judgment - **Speculative note:** Legibility is partly audience-dependent. Score against a stated reference audience. ### C5a — Narrative transparency (new, split from C5) - **Theory:** Becker (conventions of documentation and provenance); Rogers (observability) - **Measures:** Public documentation of artistic intent, datasets used (sources and nature), and rationale for key creative decisions. - **Assessment:** Checklist: intent stated publicly; datasets named and described; decision rationale available to non-specialists; public discussion of process in interviews, essays, or wall text. Score by completeness. - **Type:** Mixed - **Speculative note:** Narrative transparency may be easier to perform than technical transparency, and may function as curatorial framing rather than verifiable disclosure. ### C5b — Technical transparency (new, split from C5) - **Theory:** Becker (conventions of documentation); Rogers (observability, trialability for practitioners) - **Measures:** Archived prompts, open or documented pipeline, reproducible iteration history, third-party verification possible. - **Assessment:** Weighted checklist: model/version named; prompts archived; pipeline documented; iteration record kept; data provenance stated; third-party reproduction possible. Score by completeness and verifiability, not by claims. - **Type:** Mixed - **Speculative note:** The claim that transparency causes faster institutional attention is speculative. An internal scan of four artists (Kishi, Rothberg, Sills, Andrew) found faster attention among process-disclosing artists, but n=4 is too small. Technical transparency may be a symptom of institutional fluency rather than a cause. Treat as hypothesis under test. C5a and C5b should be tested separately, since they may diverge (see Anadol retrospective). ### C6 — Legitimation sequence - **Theory:** Becker (legitimation through institutional gatekeeping; sequence of consecration) - **Measures:** Order and tier of institutional recognition: acquisitions, group shows, solo shows, critical essays, prizes, residencies, public commissions. - **Assessment:** Map recognition events on timeline; score tier of each institution and whether sequence follows established escalation pattern. Skipping tiers scores lower unless justified. - **Type:** Evidence-based - **Speculative note:** Sequence-based legitimation is documented. Whether AI-specific sequences differ or are compressed by platform visibility is open. - **Weighting:** Most heavily weighted criterion. Supported by retrospective test (see below). ### C7 — Distribution and market infrastructure - **Theory:** Becker (distribution networks); Rogers (adoption by early adopters with resources) - **Measures:** Gallery representation, edition practices, secondary-market activity, presence across multiple cities/markets. - **Assessment:** [Incomplete in source] ## Weighting - **C6 (Legitimation sequence):** Highest weight. Retrospective test scored it most predictive. - **Other criteria:** Weights not yet set. Do not finalise weighting until C3 and C5 revisions are tested on at least one further case. ## Retrospective Test Findings ### Refik Anadol | Criterion | Pre-breakout score (2017–2019) | Notes | |---|---|---| | C3 Convention compatibility | 2/5 | Work sat between immersive media, architectural projection, and fine art. Fit poorly with gallery conventions. Subsequently broke through to MoMA. Likely a false negative (see C3 revision). | | C5 Process transparency (pre-split) | 3/5 (low confidence) | Narrative transparency (public discussion of datasets and intent) diverged from technical transparency (pipeline, reproducibility). Now scored as C5a and C5b. | | C6 Legitimation sequence | 4/5 | Cumulative path from public commissions and immersive venues to museum acquisition matched predicted sequence. | **Key finding:** The rubric would have underestimated Anadol, primarily through C3. This indicates the framework is conservative and may miss artists genuinely expanding the medium. **Limitation:** C5a and C5b were not scored separately for Anadol. The split is a revision, not a retroactive score. ## Revision Log 1. **C3:** Added productive-friction scoring note and sub-criterion C3b to distinguish weak fit from expansive fit. 2. **C5:** Split into C5a (narrative transparency) and C5b (technical transparency), to be scored separately. 3. **C6:** Retained as most heavily weighted criterion; retrospective supports robustness. 4. **C7:** Still incomplete. Assessment to be drafted. ## Overall Validity Assessment - The rubric is **conservative**: it may under-rank artists who are expanding the medium, as shown by the C3 false negative for Anadol. - C6 is the most robust criterion on current evidence. - C3 and C5 should be revised (done above) and tested on further retrospective cases before the rubric is applied to emerging artists. - Single-case evidence (n=1 retrospective) is insufficient to validate any revision. Treat all changes as provisional. ## Application to Emerging Artists (October 2026) ### Scores | Artist | C1 | C2 | C3 | C4 | C5a | C5b | C6 | C7 | Total | |--------|----|----|----|----|-----|-----|----|----|-------| | Sasha Stiles | 5 | 4 | 4 | 5 | 4 | 3 | 5 | 4 | **34** | | Gretchen Andrew | 4 | 3 | 4 | 4 | 3 | 3 | 4 | 3 | **28** | | Sarah Rothberg | 3 | 2 | 3 | 3 | 3 | 3 | 3 | 2 | **22** | | Yuma Kishi | 3 | 2 | 3 | 3 | 3 | 2 | 3 | 2 | **21** | | Victoria Sills | 2 | 2 | 2 | 4 | 3 | 2 | 2 | 2 | **19** | ### False-Negative Risk Flags - **Victoria Sills: HIGH** — C3 scored 2 because AI-to-hand-painting hybrid fits existing categories poorly and shows no documented boundary-expansion recognition yet. Friction may signal expansion rather than weakness. Low total reflects early-stage visibility, not weak work. - **Gretchen Andrew: medium** — Facetune Portraits sits at ambiguous medium boundary; C3 may under-score it. - **Sasha Stiles: medium** — Poetics-based practice may register low on medium-based AI signals if judged by C3 alone. - **Sarah Rothberg: medium** — Conversation-as-work may fall outside C3 categories. - **Yuma Kishi: low-medium** — Self-built tooling is strong C5b candidate, but C3 reading depends on medium definition. ### Key Observations - C6 (legitimation sequence) remains the most predictive criterion. Stiles and Andrew score highest largely because they have documented institutional acquisitions. - The rubric is **conservative**: it rewards artists who have already been validated and may penalize boundary-expanders who have not yet reached institutional tiers. - C3 false-negative safeguard is partially effective but may still under-score artists whose work genuinely expands medium boundaries. - C5b (technical transparency) is the weakest-scored dimension across all five artists, suggesting that open pipeline documentation is rare even among process-disclosing practitioners. ### Confidence Low to medium. Scores rely on brief-stated claims and web-search evidence. No primary-source verification of acquisition dates or exhibition scope. The rubric has been calibrated against only one retrospective case (Anadol).