Guides
Plain-English, source-linked guides to the EU AI Act's transparency rules. Not legal advice.
Start here
Does the AI Act apply to chatbots, AI content and deepfakes?
12 plain answers to the questions teams actually search — chatbots, deepfakes, AI text, non-EU companies, internal tools, open source.
Article 50 is now in force: who must do what
Paragraph by paragraph: the transparency obligations that have applied since 2 August 2026, who they cover, and the fastest way to comply.
EU AI Act Article 50, explained clause by clause
What each paragraph of Article 50 actually requires from providers and deployers — with the exemptions.
The Article 50 compliance checklist (August 2026)
A practical, step-by-step checklist to get chatbots and AI-generated content compliant with Article 50, in application since 2 August 2026.
Article 50 is in force: 5 things to do this week
In force since 2 August 2026. Confirm your obligations, ship chatbot disclosure, start machine-readable marking, label deep fakes and document it all — this week.
Was the EU AI Act delayed? What has applied since 2 August 2026
The staged Article 113 timeline, what the Digital Omnibus regulation actually changed, and why Article 50 stands.
EU AI Act: what was delayed — and what was not
The Digital Omnibus moved the high-risk rules to 2027/28 — it did not move Article 50. The full date table with EUR-Lex sources.
EU AI Act risk categories: which one is my product in?
Prohibited, high-risk, transparency and minimal risk explained with SaaS examples — and why most products land in the Article 50 tier.
By role & product
EU AI Act by role & industry: 24 guides for providers and deployers
Healthcare, HR, education, finance, insurance, e-commerce, marketing, legal, manufacturing, SaaS, media, public sector — your high-risk triggers, Article 50 duties and post-Omnibus dates.
Article 50 by product type: chatbots, generators, deep-fake tools
The provider vs deployer duties each product category triggers — disclosure, machine-readable marking and visible labelling, sorted.
Is my chatbot covered by EU AI Act Article 50?
The Article 50(1) test for support bots, sales bots and voice agents — and what the 'obvious to the user' exemption really covers.
AI chatbot disclosure examples: wording that satisfies Article 50
Copy-paste disclosure wording for chat widgets, messaging bots and voice agents — with examples that pass and formulations that fail.
Do AI voice agents and phone bots need disclosure?
Why voice is squarely in Article 50(1), inbound vs outbound scripts, and when cloned voices trigger the deep-fake rules.
Do I need to label AI-generated blog posts and marketing content?
Article 50(2) marking vs 50(4) disclosure, the human-review exemption for editorial content, and rules for deep fakes in ads.
Does the EU AI Act apply to US and other non-EU companies?
Article 2's extraterritorial reach: when non-EU SaaS companies must comply, and what 'output used in the Union' means.
GPAI obligations under the EU AI Act: Articles 53–55 explained
Documentation, copyright policy, training-data summary, the 10^25 FLOPs systemic-risk threshold — and the Commission's enforcement powers over model providers, live since 2 August 2026.
Machine-readable marking & C2PA
Machine-readable marking: the technical implementation checklist
C2PA Content Credentials, invisible watermarks and file metadata compared per content type — what to actually build before 2 December 2026.
Countdown: the Article 50(2) marking grace period ends 2 December 2026
Systems on the market before 2 August 2026 get until 2 December 2026 to add machine-readable marking. Who qualifies, what to do, and an email reminder.
C2PA & Content Credentials: machine-readable AI marking
How to satisfy Article 50(2)'s machine-readable marking requirement using the C2PA open standard.
C2PA sign image API: Content Credentials with one POST
Add signed, machine-readable AI marking to generated images via a REST API — with a free playground to try it first.
Code of Practice on Transparency of AI-generated Content: sign or not?
The Commission-backed voluntary code for Article 50(2)/(4): what Sections 1 and 2 require, what signing proves, and what the signatory list means for procurement.
Enforcement, penalties & the law
EU AI Act penalties: what you actually risk
The fine structure of Article 99, who enforces it, and why transparency breaches cost up to €15M or 3% of turnover.
EU AI Act fines calculator: what's your maximum exposure?
Interactive calculator for the Article 99 tiers — €35M/7%, €15M/3%, €7.5M/1% — plus the SME cap and how fines are really set.
Who enforces the AI Act in each country — and how complaints work
DE, FR, ES, IT, PL, NL and the EU level: the competent authorities, designation status, and how Article 85 complaints work — official sources only.
GDPR vs AI Act: what changes for chatbot disclosure?
Why a GDPR privacy notice does not satisfy Article 50(1) — the two regimes compared side by side, and how they stack on one chatbot.
Policies, evidence & procurement
AI transparency policy template: structure + free skeleton
The 9 sections an Article 50 policy must contain, a copy-paste skeleton, and the mistakes that sink policies in audits.
The EU AI Act evidence pack: prove compliance, don't just claim it
What an Article 50 evidence pack should contain — and how to export a dated, reviewable one in a click.
AI vendor due diligence: the questions to ask before you buy
You deploy your vendors' AI under your own Article 50 obligations — the questionnaire, contract clauses and red flags.
AI Act Kit vs manual compliance: an honest comparison
Tooling vs lawyer-and-DIY for Article 50 — costs, coverage, and the cases where manual is genuinely better.
For agencies & consultants
Selling Article 50 gap assessments: a consultancy playbook
How agencies and compliance consultants package gap assessments as a productised service — scoping, white-label reports, pricing, honest limits.
White-label AI Act compliance reports for agencies & consultants
Deliver gap reports and evidence packs under your own logo, name and brand colour — with traceable report numbers your clients can verify.