U.S. senators revive AI Labeling Act for images, video, and audio
Senators Brian Schatz, John Curtis, and Mark Warner introduced the AI Labeling Act of 2026, which would require clear labels on AI-generated digital content. The proposal covers AI-generated image, video, and audio content and calls for both visible and machine-readable disclosures.
Why visual creators should care
For photographers, designers, AI image creators, agencies, stock contributors, and brand teams, this is a sign that AI disclosure may become a formal compliance layer rather than a voluntary caption choice. If adopted, AI-generated campaign visuals, synthetic product images, AI social assets, and altered photorealistic content could need persistent disclosures that travel with the file or platform post.
Practical takeaway
Start treating AI disclosure as part of the production record. Keep notes on which assets were AI-generated or materially AI-altered, which tool was used, whether the content includes realistic people or products, and whether visible or machine-readable labels were applied before publication.
What to watch next
Watch whether the bill advances and whether the proposed working group pushes platforms toward interoperable disclosure, provenance, or C2PA-style technical standards for AI-generated visual content.
Sources
TikTok Shop rules tell sellers to label AI-generated or altered content
TikTok Shop Seller University says AI-generated content is allowed, but sellers must be transparent and truthful. The guidance says sellers should not use AI to mislead users about products, people, or facts, and should clearly label content that is fully or significantly created or altered by AI.
Why visual creators should care
For ecommerce sellers, product photographers, agencies, TikTok Shop creators, and product-content teams, this turns AI-generated product imagery into a platform-policy issue. AI product shots, AI UGC-style videos, synthetic model content, altered demonstrations, and AI-enhanced product claims can create risk if they misrepresent what is actually being sold.
Practical takeaway
For TikTok Shop campaigns, keep a record of which product images, videos, backgrounds, models, voiceovers, or demos were AI-generated or AI-altered. Client briefs should say who applies labels, who checks product accuracy, and who is responsible if AI content creates a false impression.
What to watch next
Watch whether TikTok Shop starts enforcing these rules more visibly through listing removals, seller warnings, creator penalties, or AI-generated-content toggles inside campaign publishing tools.
Sources
Australia copyright fight escalates around AI training and data-centres
The Guardian reported renewed concern from Australian creatives over claims that a large data-centre proposal could be linked to weakening copyright restrictions for AI training. The report says creators and politicians are pushing back against any deal that would give AI companies broader access to copyrighted material in exchange for investment or compensation schemes.
Why visual creators should care
This is a material update to the Australia AI-training story because it puts copyright policy, creator compensation, and AI infrastructure into the same negotiation. For visual creators, archives, photographers, illustrators, and agencies, the key issue is whether copyrighted images remain permission-based training material or become part of a broader policy carve-out.
Practical takeaway
For commissioned and archived visual work, keep AI-training restrictions explicit. A client licence to use images in a campaign should not automatically become permission to train AI systems, build datasets, or contribute to future model development.
What to watch next
Watch whether Australia continues to reject a broad text-and-data-mining exception, moves toward a licensing or compensation framework, or separates visual works from other types of creative material in any final proposal.
Sources
AI influencer accounts show how easily likeness can become a synthetic product
Vox reported on convincing AI-generated influencer and thirst-trap accounts that look like real people, attract large followings, and can divert attention or income from real creators. The report also describes real performers and models seeing AI versions or near-imitations of themselves used to sell attention, subscriptions, or fake relationships.
Why visual creators should care
For photographers, influencers, portrait clients, models, agencies, and visual creators, this is a likeness-rights warning. A small number of public photos may now be enough to help create a believable synthetic persona, and those personas can compete with, impersonate, or exploit the real people whose images helped make them convincing.
Practical takeaway
For work involving identifiable people, separate ordinary image use from AI likeness reuse. Model releases, influencer agreements, and client approvals should say whether faces, bodies, poses, source photos, or footage may be used to create synthetic personas, digital doubles, or AI-generated future content.
What to watch next
Watch whether platforms create stronger takedown paths for AI persona impersonation, especially where fake accounts monetise a person’s appearance, imply a relationship, or compete directly with real creator income.
Sources
Documentary photo grants and contests keep drawing a hard line on AI images
The 2026 W. Eugene Smith Grant application rules say applicants may submit representative images for review, but no AI-generated imagery is allowed. World Press Photo’s 2026 contest rules similarly state that AI-generated images are not photography and that synthetic images or generative fill will automatically disqualify an entry.
Why visual creators should care
For photographers, documentary creators, photojournalists, contest entrants, and grant applicants, the line between AI assistance and disqualifying generation is becoming more explicit. These rules show that major photography contexts still treat camera capture, source files, and edit integrity as central to trust.
Practical takeaway
Before submitting to grants, contests, or editorial calls, read the AI rules and preserve RAW files, edit histories, captions, and source records. Do not assume that AI enhancement, generative fill, background replacement, or synthetic reconstruction will be treated the same way across competitions.
What to watch next
Watch whether more grants and contests require AI-use declarations, RAW-file checks, provenance records, edit-history review, or separate categories for AI-assisted work.
Sources
Vietnam’s AI content rules bring labels and training-rights limits into focus
Vietnamese legal and official media coverage says new rules require labels for AI-generated or AI-modified content when it imitates real people or recreates real-world events in a way that could be mistaken for authentic content. Separate copyright coverage also highlights stricter requirements around text and data mining for AI training, including lawful access and respect for copyright holders’ reserved rights.
Why visual creators should care
For agencies, publishers, stock contributors, AI image creators, and international clients, this shows that AI disclosure and training-rights rules are not only U.S. or EU issues. Visual campaigns that cross markets may need to account for local rules on synthetic people, recreated events, AI labels, and whether copyrighted material can be used for training.
Practical takeaway
For international work, avoid using one generic AI-use clause everywhere. Contracts should separate AI-generated content labels, real-person likeness consent, recreated-event risk, source-image rights, and AI-training restrictions by territory where needed.
What to watch next
Watch whether Vietnam issues more detailed technical standards for labelling AI-generated images and whether similar training-data restrictions appear in neighbouring markets.
Sources
