← All work signals
Labour signal

The Model Union

Writers, artists and coders organise over the work that trained the machine, bargaining for a cut of their own reflection.

What is it?

Writers and artists are turning lawsuits into standing contract terms — a per-work settlement, guild disclosure rules, licensing clauses that survive any single case. Individual claims are consolidating into collective leverage, the same lever labour has always pulled. Whether that leverage reaches past organised guilds is unproven.

Why now?

Law never anticipated a machine that reads a career’s output as training fuel, so courts are writing the rule after the fact — catch-up, not design. Once one settlement puts a price on a work, every guild has a number to cite. Precedent is standing in for policy.

Who’s building it?

Guilds and unions negotiate the terms directly with studios and labels — labour setting its own price.

Writers' guildsMusicians' unionsLabels & studiosAI model makersCourts & settlements
The full read
Why now, who is building it, where the market goes, and what stays unsettled.
Get weekly →
01 · The market
Who’s in it
Builders — guilds and unions bargaining collectively on precedent.
Buyers — studios, labels and publishers buying licensed rights instead of risking further litigation.
Gatekeepers — courts and settlement terms, setting the price everyone else cites.
What’s at stake
The settlement that priced one work sets a going rate — a number every future claim now measures itself against.
Compliance cost lands on every studio, paid or not.
Workers outside organised guilds get no seat at all.
What’s unsettled
Whether this one settlement price eventually becomes a floor or ceiling.
Whether guild terms ever extend past entertainment into other industries.
Whether licensing becomes a routine, standing revenue line, or stays one-off settlement payouts.
02 · How the demand matureda progression, not a timeline
Litigation
Individual lawsuits test whether training on protected work is fair use.
Settlement
One case prices a work, giving every future claim a number.
Contract terms
Guilds write disclosure and consent clauses directly into standing agreements, ahead of any statute.
Standing licensing
Recurring royalty deals replace one-time payouts as the ordinary cost of business.
03 · Where the market goesthe forks, not a prediction
→Consolidation or spread — whether large, well-lawyered guilds capture the terms alone, or licensing norms spread to freelance and gig creators too.
→Studios may end up writing the real compliance rulebook before any regulator drafts a statute.
→Value may concentrate with whichever guild negotiated first, not with the workers behind the actual catalogue.
→Most global labour without a guild gets none of the leverage this creates.
The demand question underneath every clause: does this leverage widen past organised guilds to the far larger, unorganised labour force training these systems for free?
04 · The Compliance-as-Feature deskadjacent markets, moving now
The Training Consent Wars
Where the litigation over training began.
Anthropic’s $1.5B author settlement won final approval.
TechCrunch
The Explainability Demand
Regulators now require AI content to disclose itself.
EU AI Act Article 50 mandates AI-content transparency labelling.
European Commission
Consent Fatigue
Users grow weary of endless AI data consent.
LinkedIn began using member data to train AI models.
LinkedIn
05 · The Wire
01
Anthropic’s author settlement pays $3,000 per work
TechCrunch ↗
02
SAG-AFTRA members ratify four-year contract limiting AI performers
Associated Press via WTOP ↗
03
Musicians union sues labels over uncompensated AI training use
Paste Magazine ↗
The panel on this signal
They agreed on
–Litigation is steadily hardening into standing licensing terms, with guilds negotiating disclosure and consent directly into every contract.
–One large settlement now sets the going market rate.
–Organised guilds set the real terms first, well ahead of any formal law.
The labour economist disagreed on
–Settlement-driven bargaining leverage in this instance is confined to organised guilds with existing collective agreements; the much larger pool of unorganised, globally distributed contributors to training data has no comparable negotiating structure.
Know more about the method →
✦How this signal was built
01 · Panel
A Delphi panel of 4 AI analysts, each reasoning independently.
→
02 · Debate
They challenge and revise each other’s reads over two rounds.
→
03 · Evidence
Every claim is tied to a cited, linked source on The Wire.
→
04 · Synthesis
A moderator writes it into one read, disagreement and all.