borghei/Claude-Skills / marketing/ai-content-disclosure
ai-content-disclosure
VerifiedCheck AI-generated marketing content and reviews for required disclosures under the EU AI Act, FTC rules and platform AI-label policies. Use when labelling AI ads, deepfakes, chatbots, influencer posts or testimonials.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 15.4 KB |
| assets/content_manifest_template.json | config | 2.5 KB |
| assets/disclosure-label-templates.md | markdown | 3.4 KB |
| assets/sample_content_manifest.json | config | 2.7 KB |
| assets/sample_reviews.json | config | 2.6 KB |
| references/disclosure-decision-tree.md | markdown | 3.9 KB |
| references/disclosure-wording-library.md | markdown | 4.4 KB |
| references/platform-ai-label-policies.md | markdown | 5.2 KB |
| references/regulation-summaries.md | markdown | 13.6 KB |
| scripts/disclosure_checker.py | script | 16.4 KB |
| scripts/review_authenticity_linter.py | script | 10.7 KB |
references/regulation-summaries.md
View on GitHub# Regulation Summaries — AI Content and Endorsement Disclosure
Status: checked against official sources in **September 2026**. Laws, guidance and dates move; re-check the linked primary source before relying on any line here. This is decision support for marketers, **not legal advice**.
---
## 1. EU AI Act — Article 50 transparency obligations
**Instrument:** Regulation (EU) 2024/1689 (Artificial Intelligence Act), as amended by the Digital Omnibus on AI (in force 27 July 2026).
**Primary sources:**
- Regulation text: https://eur-lex.europa.eu/eli/reg/2024/1689/oj
- Commission FAQ on Article 50: https://digital-strategy.ec.europa.eu/en/faqs/transparency-obligations-under-article-50-ai-act
- Code of Practice on Transparency of AI-generated Content: https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content
- AI Act overview and timeline: https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
### Application date
| Obligation | Applies from |
|------------|--------------|
| Article 50 (all paragraphs) — general application | **2 August 2026** |
| Article 50(2) machine-readable marking — AI systems placed on the market **before** 2 August 2026 | **2 December 2026** (Digital Omnibus transitional period) |
The Omnibus transitional period covers **only** the Article 50(2) marking duty for legacy systems. Chatbot notices (50(1)) and deployer labelling (50(3), 50(4)) were not postponed.
### Who owes what
| Para | Duty holder | What it requires | Marketing examples |
|------|-------------|------------------|--------------------|
| 50(1) | **Provider** of an AI system intended to interact directly with people | Design it so people are informed they are interacting with an AI system, unless that is obvious to a reasonably well-informed, observant and circumspect person | Website sales assistant, support chatbot, AI voice agent for outbound calls |
| 50(2) | **Provider** of a system generating synthetic audio, image, video or text | Outputs marked in a machine-readable format and detectable as artificially generated or manipulated. Does not apply where the AI performs an assistive function for standard editing or does not substantially alter the input | Only relevant if you **offer** a generator to others (e.g. an in-product "create your ad" tool) |
| 50(3) | **Deployer** of emotion recognition or biometric categorisation | Inform exposed people of the system's operation | In-store camera analytics, "mood-based" ad screens |
| 50(4) first subpara | **Deployer** of a system generating or manipulating image, audio or video that constitutes a **deep fake** | Disclose that the content has been artificially generated or manipulated | Realistic AI video of a place, event or person in a campaign |
| 50(4) second subpara | **Deployer** of a system generating or manipulating **text published to inform the public on matters of public interest** | Disclose that the text has been artificially generated or manipulated | AI-written news-style article, policy explainer, health or finance commentary |
| 50(5) | Both | Information given clearly and distinguishably, at the latest at the time of the first interaction or exposure, and conforming to accessibility requirements | Label on the asset itself, not buried in a footer |
**"Deep fake"** (Art. 3(60)): AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful. It is not limited to people — a realistic fake street scene or product demo can qualify.
### Exceptions that matter to marketers
| Exception | Scope | What to keep on file |
|-----------|-------|----------------------|
| Evidently artistic, creative, satirical, fictional or analogous work | Deepfake disclosure is limited to disclosing the existence of such content in an appropriate manner that does not hamper display or enjoyment of the work | Rationale for why the work is evidently artistic/satirical; where the disclosure sits (e.g. end card, caption) |
| Human review / editorial control (text only) | 50(4) text duty does not apply where AI-generated text has undergone human review or editorial control **and** a natural or legal person holds editorial responsibility for publication | Reviewer name, date, what was changed, who holds editorial responsibility. The Commission FAQ describes editorial control as substantive examination by someone with authority to approve, alter or reject |
| Assistive function / standard editing | Removes the 50(2) marking duty where AI only assists with standard editing or does not substantially alter input | Description of the edit (spell check, colour correction, background clean-up) |
| Law-enforcement authorisation | Not relevant to marketing | — |
A **commercial ad is not automatically "a matter of public interest"** — the text duty targets content that informs the public (news-like, policy, health, safety, finance, elections). An AI-written product description is normally outside 50(4) second subparagraph; a sponsored "industry report" on new regulation is plausibly inside it. When unsure, label or document editorial control.
### Penalties
Article 99(4): non-compliance with Article 50 obligations can draw administrative fines of up to **EUR 15 million or 3% of total worldwide annual turnover**, whichever is higher (for SMEs and start-ups, whichever is lower — Art. 99(6)). National market surveillance authorities enforce.
### Code of Practice (voluntary)
Published by the Commission in June 2026. Section 1 covers provider marking and detection; Section 2 covers deployer labelling of deepfakes and AI-generated text. Signing is voluntary; it is a practical route to demonstrate compliance, not a legal safe harbour. Check the policy page above for the current signatory list and label conventions.
---
## 2. US FTC — Rule on the Use of Consumer Reviews and Testimonials (16 CFR Part 465)
**Primary sources:**
- Rule text: https://www.ecfr.gov/current/title-16/chapter-I/subchapter-D/part-465
- FTC Q&A: https://www.ftc.gov/business-guidance/resources/consumer-reviews-testimonials-rule-questions-answers
- Announcement (Aug 2024): https://www.ftc.gov/news-events/news/press-releases/2024/08/federal-trade-commission-announces-final-rule-banning-fake-reviews-testimonials
**Effective:** 21 October 2024.
| Section | Prohibited conduct (summary) | AI angle |
|---------|------------------------------|----------|
| 465.2 | Writing, creating, selling, procuring or disseminating reviews/testimonials that misrepresent that the reviewer exists, or that they used or experienced the product, or their experience — when the business knew or should have known | The FTC names **AI-generated fake reviews** as covered ("by someone who does not exist, such as AI-generated fake reviews") |
| 465.3 | Review hijacking — repurposing reviews written for one product to appear for a substantially different product | AI-assisted catalogue merges that copy reviews across SKUs |
| 465.4 | Providing compensation or other incentives conditioned, expressly or by implication, on the review expressing a particular sentiment | "Get 20% off for a 5-star review" automations |
| 465.5 | Insider reviews/testimonials (officers, managers, employees, agents, and their relatives) without clear and conspicuous disclosure of the relationship; soliciting such reviews in some circumstances | Employees posting AI-drafted "customer" reviews |
| 465.6 | Company-controlled review or "independent" comparison websites misrepresented as independent | AI-generated "best X of 2026" affiliate sites owned by the brand |
| 465.7 | Review suppression — unjustified legal threats, intimidation, false accusations to remove negative reviews; misrepresenting that displayed reviews are all or most reviews while suppressing negative ones | Auto-moderation that hides low ratings |
| 465.8 | Selling or buying fake indicators of social media influence (bot followers, views) for commercial purposes | Bot-generated engagement |
**Incentivised honest reviews are not banned** by the Rule, provided there is no express or implied sentiment condition — but the incentive still needs disclosure under the FTC Act and Endorsement Guides.
**Penalties:** The Rule lets the FTC seek civil penalties for knowing violations. The maximum per violation under Section 5(m)(1)(A) of the FTC Act was adjusted to **USD 53,088** in January 2025 (https://www.ftc.gov/news-events/news/press-releases/2025/02/ftc-publishes-inflation-adjusted-civil-penalty-amounts-2025); the current figure is in 16 CFR 1.98 (https://www.ecfr.gov/current/title-16/chapter-I/subchapter-A/part-1/subpart-L/section-1.98). Each fake review can be a separate violation.
---
## 3. US FTC — Endorsement Guides (16 CFR Part 255)
**Primary sources:**
- Guides text: https://www.ecfr.gov/current/title-16/chapter-I/subchapter-B/part-255
- FTC's Endorsement Guides: What People Are Asking: https://www.ftc.gov/business-guidance/resources/ftcs-endorsement-guides-what-people-are-asking
- 2023 update announcement: https://www.ftc.gov/news-events/news/press-releases/2023/06/federal-trade-commission-announces-updated-advertising-guides-combat-deceptive-reviews-endorsements
Key points for AI-era marketing:
1. The 2023 revision clarified that "endorsements" include fake reviews, **virtual influencers** and tags in social media.
2. Endorsements must reflect the honest opinions and actual experience of the endorser. An AI persona cannot claim to have used a product.
3. **Material connections** (payment, free product, discounts, employment, family, affiliate commission) must be disclosed **clearly and conspicuously** — the Guides define this term; for social media the disclosure should be hard to miss (the FTC's guidance calls disclosures in descriptions, comment threads or behind "more" links easily avoidable).
4. A platform's built-in disclosure tool **might not be adequate** on its own.
5. "Thanks [Brand]" alone does not communicate a paid relationship; "#ad", "Ad:", "Paid partnership" at the start of the post generally do.
6. Each new post needs its own disclosure; viewers may not have seen earlier posts.
---
## 4. EU Unfair Commercial Practices Directive — review provisions
**Primary source:** Directive 2005/29/EC as amended by Directive (EU) 2019/2161 — https://eur-lex.europa.eu/eli/dir/2005/29/oj
- Art. 7(6): a trader giving access to consumer reviews must say whether and how it ensures reviews come from consumers who actually used or bought the product.
- Annex I (always unfair): point 11 (advertorial not clearly identified as paid), point 22 (falsely presenting oneself as a consumer), point 23b (claiming reviews are from real users without reasonable and proportionate checks), point 23c (submitting or commissioning false consumer reviews or endorsements, or misrepresenting reviews to promote products).
- Applies to AI-generated reviews exactly as to human-written fakes.
---
## 5. UK — Digital Markets, Competition and Consumers Act 2024 (fake reviews)
**Primary sources:**
- CMA fake reviews guidance (CMA208): https://assets.publishing.service.gov.uk/media/67eeb64fe9c76fa33048c790/CMA208_-_Fake_reviews_guidance.pdf
- CMA unfair commercial practices guidance (CMA207): https://www.gov.uk/government/publications/unfair-commercial-practices-cma207/unfair-commercial-practices
From **6 April 2025**, banned practices include submitting or commissioning fake reviews, **concealed incentivised reviews**, and publishing reviews in a misleading way (e.g. hiding negative reviews, inaccurate star ratings). Publishers of reviews must take reasonable and proportionate steps to prevent and remove fake reviews. The CMA can fine up to **10% of global turnover** under the direct enforcement regime.
---
## 6. New York — synthetic performer disclosure in advertising
**Primary sources:** S.8420-A / A.8887-B — https://www.nysenate.gov/legislation/bills/2025/S8420/amendment/A ; Governor's announcement (in effect June 2026): https://www.governor.ny.gov/news/governor-hochul-announces-first-nation-law-requiring-disclosure-when-advertisements-include-ai
- Persons who produce or create an advertisement must disclose when it includes an AI-generated **synthetic performer** (digitally created media that appears as a real person).
- The announcement references exceptions for audio-only ads, language translation and expressive works.
- Civil penalty reported by the State: USD 1,000 for a first violation and USD 5,000 for subsequent violations. Confirm the enacted text for knowledge standards and exceptions before relying on them.
---
## 7. India — IT Rules amendment on synthetically generated information
**Primary source:** Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026 (G.S.R. 120(E), 10 February 2026; in force 20 February 2026) — https://www.meity.gov.in/static/uploads/2026/02/550681ab908f8afb135b0ad42816a1c9.pdf
- Obligations fall mainly on **intermediaries** (platforms): labelling of synthetically generated information, technical measures, takedown timelines.
- For advertisers the practical effect is platform-side: expect platforms to require or auto-apply visible AI labels on synthetic visual/audio content shown in India. Google Ads cites India among the jurisdictions requiring AI labels in ads (see platform-ai-label-policies.md).
---
## What this skill deliberately does not cover
- Political advertising regimes (EU Regulation 2024/900 on political advertising, US state election deepfake laws) beyond the platform checkbox requirements — route to counsel.
- Copyright and personality/publicity rights in AI-generated likenesses — separate clearance process.
- Sector rules (financial promotions, medicines advertising) that add their own disclosure duties.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirements | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |