WPP and Meta test AI creative tools that learn from brand assets
WPP announced that it is the first advertising partner to pilot Meta's newest AI creative solution inside WPP Open. The system is designed to use campaign performance data, a brand's existing advertising, and a brand-memory layer to generate and test new creative concepts.
Why visual creators should care
For agencies, photographers, designers, art directors, and brand teams, this shows AI creative generation moving deeper into normal campaign workflows. Existing ads, brand assets, tone, visual identity, and campaign performance data may increasingly be used as inputs for future AI-generated creative, which raises questions about client approval, asset reuse, brand control, and production records.
Practical takeaway
Client agreements should say whether approved assets, past campaigns, product images, style guides, and performance data can be used to train or guide AI creative systems. Agencies should also keep records showing which assets were used as inputs and who approved generated campaign concepts.
What to watch next
Watch whether WPP, Meta, and major advertisers publish clearer rules around brand-memory systems, asset permissions, human review, and whether creators can restrict their delivered work from being used in future AI creative generation.
Sources
Adobe's Topaz Labs deal brings AI enhancement closer to core creative workflows
Adobe announced that it has entered a definitive agreement to acquire Topaz Labs, whose AI tools are used for upscaling, sharpening, stabilising, restoring, and increasing the resolution of images and video. Adobe says Topaz technology will be integrated into Firefly, Firefly Services, and Creative Cloud apps, while Topaz tools will also continue as standalone products.
Why visual creators should care
For photographers, retouchers, videographers, agencies, and stock contributors, this is another sign that AI enhancement is becoming part of the default professional workflow. The practical distinction between ordinary enhancement, restoration, generative alteration, and factual manipulation will become more important as these tools sit closer to Photoshop, Lightroom, Premiere, and enterprise image pipelines.
Practical takeaway
Creators should record when AI enhancement tools are used, especially for editorial, documentary, stock, contest, or client-delivery work. Contracts and briefs should distinguish acceptable enhancement from generative alteration, factual change, or undisclosed image reconstruction.
What to watch next
Watch how Adobe labels Topaz-powered features inside Creative Cloud and Firefly Services, whether provenance metadata records enhancement steps, and whether stock platforms or editorial clients treat AI upscaling differently from generative image editing.
Sources
Canva's new AI terms separate generated output from licensed content
Canva's AI Product Terms took effect on June 26, 2026. The terms say users are responsible for inputs and outputs, need the necessary rights and permissions for uploaded material, may use AI outputs commercially at their own risk, and do not automatically own outputs that modify or incorporate Canva licensed content.
Why visual creators should care
This matters for designers, social-content creators, agencies, small businesses, and clients who use Canva as a production tool. A platform may allow AI commercial use, but output ownership, stock-library content, provenance metadata, uploaded inputs, and third-party rights can still sit under different rules.
Practical takeaway
Before delivering AI-assisted Canva work commercially, save the relevant AI terms, stock or library licences, uploaded source files, prompts, final exports, and client approval notes. Do not assume an AI-edited Canva library image gives the client full ownership of the output.
What to watch next
Watch whether Canva adds clearer labels, C2PA metadata controls, client-facing export records, or stricter rules around AI outputs that combine user uploads with Canva library assets.
Sources
German AI-logo ruling shows prompt-only brand work may face authorship problems
A Munich Local Court decision involving AI-generated logos dismissed copyright claims over three logo designs created with prompts. The court treated AI output as potentially protectable only where human creative influence decisively shapes the result, but found the logos in the case did not meet that threshold.
Why visual creators should care
For designers, branding studios, illustrators, and agencies, this is a practical warning around AI-generated logos and client ownership promises. A client may expect exclusive copyright in a logo, but a prompt-led AI output may not provide the same ownership position as a human-designed brand mark.
Practical takeaway
For AI-assisted logo or brand work, keep records of human design choices, sketches, edits, vector reconstruction, selection decisions, client feedback, and final manual refinements. Avoid promising copyright ownership or exclusivity for prompt-only logo outputs without checking the legal position.
What to watch next
Watch whether other European courts draw clearer lines between prompt-only outputs, AI-assisted human design, and manually refined commercial branding work.
Sources
Disney and Universal's Midjourney case keeps visual AI litigation active
The Disney and Universal copyright case against Midjourney remains active with late-June docket activity. The lawsuit targets the image generator's alleged ability to produce outputs featuring protected characters, while recent litigation activity has also put discovery and the studios' own AI use under scrutiny.
Why visual creators should care
For illustrators, concept artists, agencies, fan-art sellers, AI image creators, and brand clients, this case sits at the centre of a major visual-rights question: what happens when a generative image tool can produce recognisable protected characters, styles, or franchise-like imagery on demand.
Practical takeaway
Avoid using AI-generated visuals that resemble protected characters, entertainment franchises, logos, or branded worlds in commercial work unless rights are clear. Client briefs should also prohibit prompt instructions that request protected characters, near-copies, or franchise-style substitutes.
What to watch next
Watch discovery fights around how rights holders and AI companies use generative tools, whether the court focuses on training data, outputs, safeguards, or user prompts, and whether settlement pressure leads to licensing models for character-based image generation.
Sources
Canon's C2PA imaging system pushes provenance into newsroom workflows
Canon introduced a C2PA-compliant Authenticity Imaging System for news organisations, initially supporting selected professional camera models. The system is designed to manage image provenance records, issue certificates, and help verify content history from capture through publication workflows.
Why visual creators should care
For photojournalists, editorial photographers, agencies, publishers, documentary teams, and clients, provenance is becoming a practical trust layer. As AI-generated and AI-altered images become harder to identify by sight, source files, timestamps, content credentials, and verification certificates may become part of professional delivery.
Practical takeaway
For editorial, documentary, contest, or high-trust client work, preserve RAW files, export histories, timestamps, metadata, edit records, and any C2PA or content-credential data. Client briefs should also say whether provenance metadata must be preserved through delivery and publication.
What to watch next
Watch whether C2PA workflows move beyond newsrooms into agencies, courts, insurers, government communications, contests, stock submissions, and brand campaigns that need proof of authenticity.
Sources
