EU finalizes AI-generated content labelling code ahead of August obligations
The European Commission published the final Code of Practice on marking and labelling AI-generated content. The voluntary code is meant to help providers and deployers of generative AI systems meet AI Act transparency obligations that apply from August 2, 2026, including clear labelling in key cases such as deepfakes and AI-generated or manipulated public-interest content.
Why visual creators should care
If your images, ads, campaign visuals, social assets, or synthetic portraits circulate in the EU, disclosure expectations are becoming operational rather than optional. Clients may start asking creators and agencies to prove when a visual was AI-generated, AI-edited, or human-made.
Practical takeaway
Build AI disclosure into production notes and delivery checklists now: label synthetic or materially altered visuals where required, keep tool and prompt records, and avoid stripping provenance metadata unless the client has a clear legal reason.
What to watch next
Watch which platforms and major AI vendors sign the code and how brands update EU-facing creative, influencer, and ad-review workflows before August 2, 2026.
Sources
CREATOR Act targets commercial AI imitation of visual artists’ styles
H.R. 9112, the Creative Rights Ensuring Artists’ Technique and Originality are Reserved Act, was introduced in the House on June 2, 2026. The bill would give visual artists an exclusive right to authorize commercial exploitation or public distribution of a stylistic impersonation of that artist.
Why visual creators should care
Copyright generally protects specific works, not style by itself. This bill is notable because it is aimed directly at the commercial AI-style problem that photographers, illustrators, designers, and art directors have been raising: models or users producing market substitutes that mimic a recognizable creator identity without copying one exact image.
Practical takeaway
Do not promise clients that U.S. law already gives broad style ownership. Instead, document signature-style concerns in contracts, prohibit AI style-cloning in commissioning terms, and reserve approval rights for any AI model trained or tuned on your portfolio.
What to watch next
Watch House Judiciary activity, amendments, and whether the bill gains more cosponsors or is folded into broader AI likeness or copyright legislation.
Sources
OpenAI terms put likeness consent and public remix rights in the client-work spotlight
OpenAI’s service terms for visual capabilities say users may not reproduce a person’s likeness without express consent and all necessary rights. The terms also say that publicly sharing an image or video on the services, such as in a Sora feed or cameo context, gives OpenAI rights to operate and promote the services and gives other users limited remix rights inside the services.
Why visual creators should care
For photographers, agencies, and designers, the risk is not only whether an AI output is copyrightable. Uploading a client image, model photo, source file, or campaign video into a generative platform can trigger consent, publicity, confidentiality, and reuse issues.
Practical takeaway
Before using client or talent images in AI tools, confirm model releases, client permission, platform-sharing settings, and whether public posting would allow in-platform remixing. Keep private drafts private unless the client has approved platform-specific reuse terms.
What to watch next
Watch for updates to OpenAI’s Sora, image, cameo, and app-connection terms, especially where public feeds, remixing, third-party apps, or enterprise settings change asset-control expectations.
Sources
Adobe Stock reminds contributors that AI mislabelling can get files rejected
Adobe Stock’s contributor rejection guidance says content may be rejected, suspended, or terminated for noncompliance, including failing to check the generative AI box for AI-created submissions. The guidance also flags submissions generated with references to living copyrighted artists’ names, real known people, fictional characters, properties, and titles or keywords implying actual newsworthy events.
Why visual creators should care
This is a concrete stock-marketplace rule set, not an abstract policy debate. Contributors who treat AI outputs like ordinary illustrations without disclosure can face rejection or account risk, while buyers rely on accurate labels to assess licensing and reputational exposure.
Practical takeaway
For stock submissions, save prompts and tool records, mark generative AI truthfully, remove living-artist style references from prompts/titles/keywords, and avoid implying that synthetic scenes depict real events.
What to watch next
Watch whether Adobe, Getty/iStock, Shutterstock, and other marketplaces tighten enforcement or add stronger audit trails for AI-generated and AI-edited submissions.
Sources
Adobe AEM adds Content Credentials visibility inside asset workflows
Adobe Experience Manager Assets now supports Content Credentials in the Assets view, letting users inspect details such as whether an image was AI-generated or edited, which AI tool may have been used, and process history. Adobe says the credentials remain intact when an asset is downloaded, published, or shared.
Why visual creators should care
For agencies and in-house creative teams, provenance is becoming part of asset management. If clients ask whether a final visual used Firefly or another AI tool, the answer may need to live in the asset record, not only in a producer’s email.
Practical takeaway
Use Content Credentials or similar metadata as part of final delivery for AI-assisted work, and decide in advance who is responsible for preserving, inspecting, or removing metadata during export and publication.
What to watch next
Watch whether more DAM, CMS, ad-tech, and stock systems make provenance metadata searchable, mandatory, or tied to approval status.
Sources
UK committee says broad AI-training copyright exception is no longer preferred
The UK Parliament Communications and Digital Committee published its response to the government’s AI and copyright position, noting that the government no longer prefers a broad copyright exception for AI training with an opt-out mechanism. The committee chair urged statutory transparency requirements for large AI developers and warned that best practice alone will not support licensing or enforcement.
Why visual creators should care
The UK remains a key market for agencies, photographers, illustrators, stock libraries, and rights-management organizations. Whether the UK chooses licensing, opt-out, transparency, or a new exception will affect how creators control training uses of portfolios and archives.
Practical takeaway
For UK-facing contracts, keep AI training rights separate from ordinary promotional, editorial, or commercial-use licenses. If you license images to clients, spell out whether machine-learning training, model tuning, synthetic-data generation, or style-model creation is allowed.
What to watch next
Watch the UK government’s next consultation steps, any proposed transparency framework, and whether large AI developers face mandatory source-disclosure requirements.
Sources
