Public freshness record
AI disclosure rule and platform-policy changelog
This record separates changes to law, official guidance, platform policy and source reviews from product releases. Each entry shows the reviewed sources, operational impact, affected pages and review attribution. A source recheck with no material change is labelled as such.
Material-change publication bundle
01
Updated source
Dates, status and the reviewed snapshot marker are refreshed.
02
Impact review
Dependent interpretations and rules are mapped; affected pages are revised after editorial review.
03
Public explanation
The changelog records impact; significant changes may also receive a short analysis.
Dated updates
Philippines ASC 2026-014A: materiality plus documentation dependencies
The checker separates affected AI elements, their underlying documentation requirements and current-use document clearance. A material AI change alone no longer determines the Philippine ASC disclosure branch. Missing facts require review; a missing current-use document blocks publication independently of the label.
The supplement PDF is dated October 1 and its page was published October 2, without a separate new effective date. October 2 is Disclara's editorial cutover; October 1 remains a date-review interval. The real-child permit example is counterfactual, not a blanket synthetic-child permit requirement. Visual font/size constraints and relevant child frames/layouts are recorded, not automatically verified. Retained decisions and other market/platform rules are preserved.
Primary and first-party sources (1)
- ASC Circular 2026-014A — Supplemental guide to AI in advertising (Ad Standards Council Philippines)
- Reviewed by
- Disclara source interpretation; not legal sign-off
- Role
- Primary-source product interpretation and technical checks
- Review date
India ASCI synthetic-advertising guidance scheduled; consent and disclosure decisions separated
The checker now evaluates the final ASCI self-regulatory guidance against the planned India publication date. It separates consumer-material AI-origin labelling from the special Sponsored by [Brand] disclosure for paid AI-generated product suggestions. A versioned synthetic-performance scope distinguishes likeness, new statements, endorsement, approved script and claim references, derivatives and revocation.
The ASCI guidance takes effect after the three-month period from its 29 September publication; Disclara applies it from 29 December 2026. An accurate native AI label may fulfil the AI-origin disclosure, but is not assumed to identify a paid sponsor. Rights scope and content are user supplied; an AI label never cures a fabricated endorsement or missing consent. Agreement validity and the final media need accountable review.
Primary and first-party sources (1)
- Guidelines for Responsible Labelling of Synthetically Generated Content in Advertising (Advertising Standards Council of India)
- Reviewed by
- Disclara Editorial
- Role
- Primary-source product interpretation; not legal sign-off
- Review date
Article 50 overview clarifies provider marking, editorial review and adapted disclosure
The overview now includes provider marking of synthetic text as well as audio, image and video. It separates the limited provider-side transition for qualifying pre-existing systems from deployer disclosure, states the substantive public-interest text review and editorial-responsibility conditions, and moves artistic or satirical deepfakes out of the non-application list.
Marketers should not infer that the 2 December 2026 provider transition delays their own publication duty. Artistic deepfakes still require appropriately presented disclosure, while grammar correction alone does not establish the public-interest text exception. This is a source-checked editorial clarification, not legal sign-off for an individual campaign.
Updated pages
Primary and first-party sources (3)
- Regulation (EU) 2024/1689 (Artificial Intelligence Act) (European Union)
- Guidelines on transparency obligations for providers and deployers of certain AI systems (European Commission)
- Transparency obligations under Article 50 of the AI Act (European Commission)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Multiple provenance claims in one disclosure check
The public checker now accepts up to 20 separate caller-supplied provenance observations in one asset workflow. Each claim records its scheme, proof type, scope, reported state, external-verifier availability and reported export/transcode behavior separately.
The evidence envelope preserves all claims and interprets each within its stated scope. No combined 'authenticated' or truth score is inferred. Disclara still does not parse or validate Apple Reference Image, SynthID or other supplied files; a reported external validation remains a report, not independent verification.
Updated pages
Analysis
Primary and first-party sources (3)
- Apple Reference Image: A New Approach for Verified Photography (Apple Security Research)
- View reference images in Photos on iPhone (Apple Support)
- C2PA Technical Specification (2.4 series) (Coalition for Content Provenance and Authenticity)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Provenance claim type separated from reported validity
The checker can record separate C2PA, Apple Reference Image, SynthID, IPTC/XMP and other-scheme claims with explicit type, scope, reported validation state and export survivability. Apple Reference Image is documented as sensor-capture evidence, not proof that the scene was genuine or the photographer identified.
Additional claims are interpreted as supplied evidence, never as independently verified by Disclara. Apple Reference Image parsing, signature validation and cross-platform interoperability remain unimplemented. Apple says Reference Image capture is unavailable in the EU at launch, while viewing imported reference images is supported on Apple OS 27 devices.
Updated pages
Analysis
Primary and first-party sources (3)
- Apple Reference Image: A New Approach for Verified Photography (Apple Security Research)
- View reference images in Photos on iPhone (Apple Support)
- Apple debuts iPhone 18 Pro and iPhone 18 Pro Max (Apple Newsroom)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Client-specific prohibitions added ahead of disclosure
Versioned, user-supplied client-policy rules can prohibit use or require consent, disclosure or legal approval for bounded AI scenarios. A prohibition blocks publication; adding a disclosure does not cure it. A hosted-model provider chain can be recorded separately from the host tool's terms.
The Stanford marketing guidance illustrates why rights, truthfulness and organisation policy need their own gate. Disclara does not import Stanford rules into other clients, authenticate supplied policies or approve a hosted model's licensing from host approval alone.
Updated pages
Analysis
Primary and first-party sources (1)
- AI guidelines for marketing and communications (Stanford University Communications)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Philippines ASC AI advertising materiality test added
ASC Circular 2026-014 distinguishes broad AI use during advertising production from a disclosure trigger for content created, generated or materially enhanced with AI. The checker now records production stage, final-ad change, claim and depiction impact, and the rationale for a materiality assessment.
For Philippine advertising planned from 21 September 2026, a declared AI-generated component or documented material enhancement triggers the ASC self-regulatory disclosure layer. Unresolved or contradictory facts route to human review. Documented non-material assistance does not automatically trigger the ASC rule; other law, platform and client-policy layers remain separate. No numerical threshold is inferred.
Analysis
Primary and first-party sources (1)
- ASC Circular 2026-014 — Guidelines on the Use of Artificial Intelligence in Advertising (Ad Standards Council Philippines)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Microsoft Ads uncertainty retained; YouTube AI use wording aligned
Microsoft's linked AI-generated advertising help page still returns an unavailable result. The current Copilot page verifies asset generation, not a universal per-asset C2PA or viewer-label capability. The Microsoft Ads guide remains uncertain. The YouTube long-tail guide now follows the documented Attributes → AI use wording and does not claim an unverified API workflow.
No format-specific Microsoft disclosure mechanism is promoted to verified. Readers are directed to inspect the delivered asset and separately assess law, policy and rights. YouTube decisions are based on the final realistic alteration rather than the mere use of a synthetic tool.
Primary and first-party sources (4)
- Guidelines for AI-generated advertising (Microsoft Advertising)
- Copilot in Microsoft Advertising (Microsoft Advertising)
- Microsoft Advertising Agreement (Microsoft Advertising)
- Disclosing altered or synthetic content — YouTube Help (YouTube)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Platform disclosure examples and provenance explainer reviewed
Facebook, Instagram, TikTok and YouTube guides now distinguish concrete image, video, voice, ad and scheduler cases. The YouTube guide follows the documented Attributes → AI use terminology. The provenance hub now explains an end-to-end asset handoff and separates file signals from truth, rights and disclosure decisions.
Readers can choose the disclosure checker for a rule decision or browser-only AI Label Preflight for image signals. Platform workflows are based on first-party documentation checked on 23 September 2026, not live tests of every UI or publishing API. Cookie-free CTA and preflight-use events are logged with bounded categorical dimensions; no cross-session attribution or analytics dashboard is provided.
Updated pages
Analysis
Primary and first-party sources (5)
- Our Approach to Labeling AI-Generated Content and Manipulated Media (Meta)
- Expanding GenAI Transparency for Meta's Ads Products (Meta)
- About AI-generated content — TikTok Support (TikTok)
- Disclosing altered or synthetic content — YouTube Help (YouTube)
- C2PA Technical Specification (2.4 series) (Coalition for Content Provenance and Authenticity)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Browser-only AI Production Stack Preflight released
A free browser tool now applies evaluator 13.1.0 to one supplied campaign snapshot: organization policy, client approval, allowed content and markets, restricted use cases, provenance capability, rights review and accountable human approval. Missing bounded content-type or market scope now fails closed to REVIEW_REQUIRED. The result includes a deterministic decision ID and downloadable JSON evidence.
Teams can test whether their own recorded policy permits a tool before running the separate disclosure assessment, then replay the downloaded JSON in the public Evidence Envelope Verifier. Inputs stay in the browser and are not retained. Verification detects changed inputs, decision identifiers, rulesets and availability boundaries, but is not vendor certification, evidence authentication or a cryptographic signature; retained profiles, monitoring and DAM connectors remain planned.
Primary and first-party sources (3)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
AI production-stack and DAM decision-provenance schema added
The checker schema can now receive a campaign-specific AI tool-policy snapshot covering organization and client approval, allowed content types and markets, restricted use cases, rights-model review, provenance capability and required human approval. A public versioned JSON Schema documents those inputs, while evidence replay verifies the tool decision and external DAM decision-provenance identifiers.
A supplied tool restriction can block publication before the disclosure assessment is treated as complete. Decision evidence can reference the external media system of record without copying its library, while incomplete, unknown or lost downstream provenance routes to review. Telemetry records categorical states only; live Bynder, Fotoware, Adobe and webhook connectors remain planned integrations.
Updated pages
Primary and first-party sources (3)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Exact platform capability matching enforced
The checker now resolves native disclosure controls only when platform, media type and publication workflow all match an effective capability entry. A tracked platform without an exact format/workflow record returns PLATFORM_ONLY; an untracked destination returns UNAVAILABLE.
A live stream, scheduler or other unsupported workflow can no longer inherit a toggle, reversibility claim or publishing instruction from a platform's native-upload record. Independent legal and platform-policy findings remain visible, while the mechanism fails closed to unknown and activates a bounded fallback when needed.
Primary and first-party sources (3)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Live disclosure fallback and evidence plan added
The checker now models live streaming as a distinct publication workflow. When a legal, platform or client layer requires or recommends disclosure but no native control is verified, the result supplies format-specific overlay or spoken fallback options and an evidence checklist for the delivered stream.
A missing platform toggle no longer ends the workflow. Teams can preserve a pre-stream preview, timestamped screenshot, recording and transcript without treating a secondary report about one platform as an authoritative policy source.
Primary and first-party sources (2)
- Transparency obligations under Article 50 of the AI Act (European Commission)
- Twitch CEO says platform does not need AI disclosure rules for now (GamesBeat)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Provider marking grace period and Article 50 paragraph mapping corrected
Provider-side checks now record whether the AI system was first placed on the EU market or put into service before 2 August 2026. The evaluator applies the limited Article 50(2) grace period until 2 December 2026, fails closed when that timing is unknown during the transition window and creates a dated recheck. Public-interest text is now mapped to Article 50(4), with Article 50(5) retained for disclosure manner and timing.
Providers no longer receive an unconditional marking result during the transition period. Pre-existing systems receive a 2 December recheck, newer systems remain in scope from 2 August, and evidence records preserve the system timing behind the decision.
Primary and first-party sources (3)
- Regulation (EU) 2024/1689 (Artificial Intelligence Act) (European Union)
- Transparency obligations under Article 50 of the AI Act (European Commission)
- Guidelines on transparency obligations for providers and deployers of certain AI systems (European Commission)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Article 50 Obligation Role Resolver added
The checker now identifies whether the supplied activity is provider-side, deployer-side, performed by an employee or contractor under organisational authority, independently professional, or genuinely personal and non-professional before evaluating Article 50. Provider machine-readable marking and deployer human-visible disclosure are returned as separate decisions.
A marketer using a third-party tool is no longer instructed to implement the provider's technical marking merely because they publish the output. Employees are mapped to the controlling organisation, unclear agency authority fails closed to review, and personal-use scope remains separate from platform and client rules.
Primary and first-party sources (3)
- Regulation (EU) 2024/1689 (Artificial Intelligence Act) (European Union)
- Transparency obligations under Article 50 of the AI Act (European Commission)
- Guidelines on transparency obligations for providers and deployers of certain AI systems (European Commission)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
No-policy and no-native-mechanism platform states added
The Platform Capability Registry can now distinguish an established policy, a verified absence of a dedicated platform disclosure policy, an unknown policy state and a non-applicable state. Native mechanisms separately support AVAILABLE, NO_NATIVE_MECHANISM, UNKNOWN and NOT_APPLICABLE.
Missing platform controls can no longer be interpreted as permission or as proof that disclosure is unnecessary. Where a platform has no verified native mechanism, the checker preserves independent legal, contractual and client-policy decisions and can direct teams to a visible or audible fallback.
Primary and first-party sources (2)
- Twitch CEO says platform does not need AI disclosure rules for now (GamesBeat)
- Transparency obligations under Article 50 of the AI Act (European Commission)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
California realism threshold and enacted exclusions added
The California SB 1050 evaluator now requires a realistic synthetic human performance and separately records whether the advertisement promotes an expressive work with a performer consistent with that work or uses AI only for translation or accessibility. Unknown exclusion facts fail closed to human review.
A stylized or otherwise non-realistic avatar no longer triggers this California rule. Covered fictional synthetic performers still receive the statutory wording constraint, while digital replicas of identifiable real people remain in the separate rights and consent path.
Primary and first-party sources (2)
- SB-1050 False advertising: synthetic performers (California Legislative Information)
- California Constitution, Article IV, Section 8 (California Legislative Information)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Disclosure durability and synthetic-performer rendering added
The checker now distinguishes platform-native, asset-visible, spoken and provenance-only disclosures, records expected survival across download, screenshot and cross-platform repost, and returns derivative disclosure risk without assuming that a native label travels with the file. Synthetic performers are now classified by identity, prominence, commercial role and audio-only use.
High-risk or likely-to-be-redistributed assets can be routed to an asset-level visible or spoken disclosure. California SB 1050 now fires only for a prominent fictional synthetic performer or generic synthetic voice in a covered advertising role from 1 January 2027, while real-person digital replicas remain a separate rights and consent case.
Primary and first-party sources (4)
- SB-1050 False advertising: synthetic performers (California Legislative Information)
- California Constitution, Article IV, Section 8 (California Legislative Information)
- Our Approach to Labeling AI-Generated Content and Manipulated Media (Meta)
- Content Credentials: Deployment Guidance (C2PA Technical Working Group)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Source Conflict Sentinel and platform-label evidence model added
The checker now versions source surfaces separately, records web/PDF hashes and fails closed when materially different versions from the same authority apply. Platform-label evidence now separates observed label state, self-disclosure, metadata signals and platform detection; an absent label is never treated as proof that content is not AI-generated.
Belgian AI advertising scenarios are routed to SOURCE_CONFLICT — HUMAN_REVIEW while the official Article 9.1 web/PDF conflict remains active. Platform capabilities now expose label reversibility, rollback confidence, post-publication edit support and label-state verifiability.
Primary and first-party sources (4)
- Belgian Advertising Code (Conseil de la Publicite / Raad voor Reclame)
- Advertising rules now centralised in one code (United Brands Association)
- AI Video Faking UK Politician's Immigration Views (Meta Oversight Board)
- Labeling AI Content — Meta Transparency Center (Meta)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Approved-versus-served creative evidence added
The checker now records the approved asset version and hash, permitted platform transformations, detected generated variants, served hash, creative diff, human re-approval requirement and disclosure state of the served variant.
Teams can distinguish the file a client approved from the creative an audience received. Legal duty, platform policy, native disclaimer capability, provenance, observed enforcement and the recommended action are now explicit independent layers; lack of rejection never lowers a requirement.
Updated pages
Primary and first-party sources (5)
- Microsoft Advertising Agreement (Microsoft Advertising)
- Guidelines for AI-generated advertising (Microsoft Advertising)
- Copilot in Microsoft Advertising (Microsoft Advertising)
- What's with the AI ads? — developer response (Timberborn developers via Reddit)
- Has anyone actually had an ad flagged or rejected? (Reddit community discussion)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Synthetic UGC representation risk added to the checker
The checker now records whether commercial content presents itself as a brand ad, testimonial, review-style post, spokesperson, creator endorsement, organic post or an unresolved format. It separately detects a synthetic performer and a claimed real product experience.
An AI label no longer ends the analysis. Synthetic testimonial or review-style content, real-experience claims and unresolved synthetic UGC can be routed to accountable human or legal review while the disclosure verdict remains a separate decision layer.
Updated pages
Primary and first-party sources (3)
- Directive 2005/29/EC — consolidated text (European Parliament and Council of the European Union)
- Guidance on the interpretation and application of Directive 2005/29/EC (European Commission)
- Stamping and Recovering Content Provenance — A Last-Mile C2PA Toolkit (IBC Accelerator Media Innovation Programme)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Publication-date rules and asset-level rights attestations added
The checker now selects known rules for the planned publication date, detects scheduled policy changes across the campaign window and records scoped likeness or voice authorization plus evidence of suspected unauthorized use.
Teams can distinguish today's check from the rules expected at launch, create a dated recheck gate and export evidence showing either the authorization scope or the recorded absence of authorization for a specific asset version.
Primary and first-party sources (4)
- Gen AI labels (Pinterest)
- Changes to our policies (Pinterest)
- IAB AI Transparency and Disclosure Framework (Interactive Advertising Bureau)
- Deepfakes are wrecking influencers’ credibility, one fake ad at a time (The Guardian)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Platform policy registry expanded and rechecked
Disclara added first-party workflow coverage for LinkedIn, Google Ads, X and WhatsApp Channels and refreshed the existing Meta, TikTok and YouTube source records.
The checker and guides can keep platform-native labels separate from EU law while exposing rollout gaps and unsupported publishing paths.
Updated pages
Primary and first-party sources (4)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
TikTok, YouTube and Meta policy review completed
TikTok, YouTube and Instagram/Meta AI-labeling policies were reviewed against first-party sources. No material decision change was identified in that review.
Existing platform guidance remained current; the review date was refreshed without implying that the underlying policy changed.
Primary and first-party sources (3)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
EU AI Act transparency obligations now apply
Article 50 obligations — including deepfake disclosure by deployers and machine-readable marking of synthetic content by providers — apply from 2 August 2026.
Providers and deployers need separate implementation paths; marketers must assess the actual content and role instead of applying one universal AI label.
Primary and first-party sources (2)
- Regulation (EU) 2024/1689 (Artificial Intelligence Act) (European Union)
- Guidelines on transparency obligations for providers and deployers of certain AI systems (European Commission)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Meta updates AI-labeling governance
Meta published its Oversight Board-informed approach to labeling AI content, including treatment of manipulated media that could mislead.
Publishers should verify the visible post-publication label and keep their own disclosure decision because technical detection is incomplete.
Primary and first-party sources (1)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
EU AI Act entered into force
Regulation (EU) 2024/1689 entered into force, with staged application dates through 2027.
Disclara tracks the application date for each relevant obligation instead of presenting entry into force as immediate application of every article.
Updated pages
Primary and first-party sources (1)
- Regulation (EU) 2024/1689 (Artificial Intelligence Act) (European Union)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
TikTok AIGC labeling requirement in effect
TikTok began requiring creators to label AI-generated content containing realistic images, audio and video.
Creators need to use TikTok’s native AI-generated-content control for covered media and retain separate evidence of the decision.
Updated pages
Primary and first-party sources (1)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
YouTube altered-content disclosure launched
YouTube introduced the altered-content disclosure requirement for realistic synthetic or altered media.
Uploaders need to answer YouTube’s altered-content question accurately and distinguish platform disclosure from the legal assessment.
Updated pages
Primary and first-party sources (2)
- Reviewed by
- Disclara Editorial
- Role
- Source and rules review
- Review date
Review process: editorial policy. Source registry: sources and methodology.