Open research / The Architect Lab / As of 28 October 2026
The Creative Infrastructure Evidence Library.
What do we actually know about the business machinery behind fashion, art and music?
A compact, source-linked research register. Each figure has a date, geography, population, original source and a question the data does not answer. Claims without context are just nice-looking numbers.
What the current evidence suggests
Different industries, different incentives.
A UK government source reports 53% AI use in design and designer fashion versus 22% in a combined group of music, performing and visual arts businesses. These are adoption measurements, not a test of value delivered.
The 2025 global recorded-music market reached $31.7bn in revenues, while 53% of UK grassroots music venues reported no profit. Those figures refer to different segments and geographies; the contrast suggests questions, not causation.
Art Basel/UBS estimates dealer operating costs rose by an average 5% in 2025 while dealer sales rose 2%. That makes margin pressure worth investigating alongside the glamour of art-market price records.
Cross-industry / source register
The baseline
51
Creative businesses using AI
Reported adoption, not successful integration or return on investment.
Unanswered: Where is AI producing measured savings rather than merely being tried?
Original source: UK Government, AI Adoption Plan: Creative Industries ↗33
All businesses using AI
Comparator supplied in the same government report; not a measure of maturity.
Unanswered: Why does creative-sector experimentation exceed broader business use?
Original source: UK Government, AI Adoption Plan: Creative Industries ↗145.8
UK creative industries gross value added
GVA is not turnover, revenue or software market size. All creative sectors included.
Unanswered: How much value is supported by invisible administrative infrastructure?
Original source: UK Department for Culture, Media and Sport, DCMS Sectors Economic Estimates GVA 2024 (provisional) ↗Fashion / source register
Fashion
53
Design and designer fashion businesses using AI
AI use includes operational and creative functions; it does not mean AI-designed collections.
Unanswered: Which functions are being used and who owns the approval?
Original source: UK Government, AI Adoption Plan: Creative Industries ↗46
Fashion executives expecting industry conditions to worsen
Sentiment survey, not a forecast of actual performance.
Unanswered: Where is pressure on margins driving operational technology investments?
Original source: BoF and McKinsey, State of Fashion 2026 ↗76
Fashion executives naming trade disruptions and duties as an industry challenge
The wording differs across executive summaries; do not mistake this for 76% of firms affected.
Unanswered: Which supplier, sourcing and approval systems are most exposed?
Original source: BoF and McKinsey, State of Fashion 2026 ↗Art / source register
Art
59.6
Global art market sales
Up 4% year-on-year. Not directly comparable to recorded music revenue.
Unanswered: Where do higher transaction values conceal operational fragility?
Original source: Art Basel and UBS Art Market Report 2026 ↗5
Growth in dealers' operating costs
Reported estimated average cost growth, faster than 2% dealer sales growth.
Unanswered: Can better operating systems relieve this margin pressure?
Original source: Art Basel and UBS Art Market Report 2026 ↗35
Art-fair share of dealers' turnover
A revenue share, not a profit share or share of all artworks.
Unanswered: How do fair logistics, client follow-ups and shipping systems scale?
Original source: Art Basel and UBS Art Market Report 2026 ↗31
Galleries viewing AI-driven discovery as a complementary aid
Self-selected survey responses. Not representative of all galleries globally.
Unanswered: What earns trust in AI recommendations without removing curator judgement?
Original source: Artsy AI Survey 2026 ↗29
Galleries concerned about AI-driven art discovery
Same sample and question as A04. Describes sentiment, not adoption.
Unanswered: Is AI discovery seen as a sales channel or a threat to mediation?
Original source: Artsy AI Survey 2026 ↗Music / source register
Music
22
Music, performing and visual arts businesses using AI
This is a combined subsector. It cannot represent music businesses alone.
Unanswered: How does adoption vary between labels, venues, performers and managers?
Original source: UK Government, AI Adoption Plan: Creative Industries ↗31.7
Global recorded-music revenues
Up 6.4% year-on-year. Not the total music economy.
Unanswered: How much growth reaches artists and operational teams?
Original source: IFPI, Global Music Report 2026 ↗53
UK grassroots music venues showing no profit
Profitability statistic, not 53% of venues closing.
Unanswered: What can administration savings realistically change at venue level?
Original source: Music Venue Trust, Annual Report 2025 ↗2.5
Average profit margin for UK grassroots music venues
Sector-level reported average; no claim that every venue has 2.5% margins.
Unanswered: Where are settlement and ticket reporting costs concentrated?
Original source: Music Venue Trust, Annual Report 2025 ↗How to use this responsibly.
This is a manually curated evidence register, not a representative survey, automated scrape or independently audited dataset. Sources are linked, not republished. Figures have different bases: revenues, GVA, market estimates, adoption rates, profitability and subjective survey answers. Do not add or rank them as though they measured the same thing.
Published dates may differ from measurement periods. The category “music, performing and visual arts” is not a music-only estimate. A survey of gallery respondents is not the global art trade. The CSV includes precise caveats and follow-up questions.
Next evidence to collect: staff hours per process; contract and rights handoffs; number of software tools; rework rates; average days to settlement; provenance verification burden; cost of administration relative to revenue; buyer/artist response times; operating budgets by firm size. These require primary interviews or client permissions, not inference from national statistics.
For a deliberately less formal perspective, try The Taste Test, an editorial pilot from The Unfamiliar Table exploring what people in art, fashion and music value when forced to choose.
For corrections, source suggestions or research collaboration: emma@thearchitectlab.com. By Emma Pulbrook.