EU lawmakers back pay-for-training: what this means for creatives
The European Parliament has called on the EU to make companies pay when they use copyrighted work to train generative AI. The goal: protect the creative sector-especially news media-whose content is being scraped, summarized, and served elsewhere without fair return.
MEPs adopted the recommendations with 460 votes in favor, 71 against, and 88 abstentions. They warn that AI-driven "answer engines" are pulling traffic away from publishers and creators, cutting into revenue and weakening media pluralism.
What the Parliament wants
- Licensing rules for AI training, including voluntary collective licensing options that cover individual creators and SMEs.
- A clear opt-out so rightsholders can exclude their work from AI training, potentially managed via an EUIPO-led list.
- No copyright protection for content fully generated by AI.
- Obligations on digital services to act against manipulated and illegal AI-generated content.
- Fair aggregation of news to protect media pluralism and prevent self-preferential practices by platform gatekeepers.
"Generative AI tools are undermining the business model of many news media organisations, by diverting click traffic away from their websites and social media channels," said MEP Barry Andrews after the vote. He added that many tech platforms now answer queries directly-"zero-click searches"-with people even talking about "Google Zero," where Google becomes an answer engine instead of a search engine.
Rapporteur Axel Voss underscored the need for legal certainty: developers must know what content can be used and how to license it, while rightsholders must be protected and paid.
Industry group CCIA Europe pushed back, arguing the non-binding report creates legal uncertainty and could slow Europe's digital competitiveness.
Why this matters if you create for a living
- Your work powers AI products that often keep users on their platforms instead of sending them to you, shrinking ad, subscription, and affiliate income.
- Licensing clarity and opt-outs could rebalance the equation-less unauthorized training, more permissioned use with payment.
- Collective licensing could give independents and small studios real leverage without hiring a legal team.
- Rules against manipulated and illegal AI content help protect your brand from impersonation and deepfake abuse.
What to do next (practical steps)
- Audit your catalog: list what you own, where it lives, and what's licensed. Add clear copyright notices and machine-readable metadata across your site and feeds.
- Set up AI training opt-outs: add robots.txt and meta tags that signal text-and-data-mining restrictions, and be ready to enroll in an EUIPO-managed opt-out list if/when it appears.
- Upgrade licensing: publish simple terms for AI use (training, embedding, summarization), with pricing tiers and a contact path for API or bulk deals.
- Track traffic shifts: monitor referral sources, zero-click exposure, and branded search drops. Adjust content formats and headlines that win actual clicks, not just mentions.
- Brand protection: watermark assets, hash your originals, and set alerts for impersonations and scraped replicas on major platforms.
- Distribution hedge: grow owned channels (newsletters, memberships) and diversify beyond a single platform or search engine.
If implemented, these measures won't stop AI-but they can make it pay you fairly and reduce the silent siphoning of your audience. Independent outlets, including those like The Brussels Times, have already felt the pain. Don't wait for a final law to start tightening your rights and routes to revenue.
For skill-building on responsible AI use in creative work, see AI for Creatives. Journalists and writers can explore AI for Writers for workflows, licensing notes, and monetization tactics.
To understand the potential role of a central opt-out list, keep an eye on the EUIPO, which MEPs suggested could manage it.
Your membership also unlocks:
