Sector guidance for common AI Act use cases

AI Act sector guides by use case help teams quickly identify the correct risk lane and obligations for their specific application. Customer chatbots and AI-generated marketing content typically trigger Article 50 transparency rules (inform users they are interacting with AI and mark generated content in a machine-readable way). Recruitment screening tools, worker-management systems, exam proctoring, and credit or insurance decision support are often high-risk under Annex III, bringing obligations around data governance, technical documentation, human oversight, and evidence collection. AI literacy requirements apply across all uses.

This hub routes you to the most relevant sector page with tailored checklists, examples, and tools focused on current law as of April 2026. Use the chooser below to match your work to the right starting point. No page promises certification or legal advice — they support readiness, evidence building, and workflow.

Law status (May 2026)

• In force now: Article 4 AI literacy obligations for providers and deployers (since 2 February 2025) and prohibited practices.
• Upcoming: Article 50 transparency obligations for chatbots, deepfakes, and certain AI-generated content apply from 2 August 2026. High-risk AI system rules (recruitment, education, credit scoring, worker management) begin in phases from August 2026 onward.
• Proposal stage: The Digital Omnibus simplification proposals and related Council/Parliament discussions in early 2026 remain under negotiation and are not current law. Always anchor in the consolidated Regulation (EU) 2024/1689 and official guidelines.

Match your AI application to the right obligations instead of reading the full text. The EU AI Act is risk-based: most interactive customer chatbots fall under transparency (Article 50), while tools that influence hiring, worker evaluation, education outcomes, or access to credit/insurance are frequently high-risk. Marketing and publishing teams using generative AI for public content must focus on disclosure and marking.

• Customer chatbots: Providers must inform users they are interacting with an AI system unless it is obvious. Deployers ensure clear communication. Link: Chatbots and the EU AI Act
• Recruitment screening: CV screening, candidate scoring, or automated shortlisting tools are typically high-risk (Annex III point 4 on employment and worker management). Requires risk management, high-quality data, technical documentation, and human oversight.
• HR and worker-management features: Performance evaluation, task allocation, or promotion recommendation tools often qualify as high-risk. Deployers must monitor operation and retain logs.
• Exam proctoring and education tools: AI used for student assessment, scoring, or proctoring can be high-risk because it affects access to education and future opportunities.
• Credit or insurance decision support: Tools that assess creditworthiness or influence insurance terms are high-risk (Annex III point 5 on access to essential services). Strong data governance and documentation are required.
• Marketing, publishers, and AI-generated public content: Generative systems producing text, images, or video for public interest matters trigger Article 50(4) disclosure obligations for deployers. Providers must enable machine-readable marking and detection. Exceptions exist for clearly artistic, satirical, or human-reviewed editorial content. Link: Marketing agencies, publishers, and AI-generated public content

These sector pages use the language of the actual team responsible — recruiters, marketers, compliance leads in education or finance — and link directly to practical artifacts.

Example: A bank deploys a generative AI chatbot that also scores credit applications. The conversational part triggers Article 50 transparency (clearly inform users it is AI). The credit-scoring component is high-risk, requiring the deployer to follow instructions, perform human oversight, and keep logs. The sector pages help you separate these obligations by role (provider vs. deployer) and give you the exact evidence checklist for each.

Teams searching “AI Act HR chatbot recruitment marketing” want answers in the language of their daily work, not abstract legal chapters. These pages connect immediately to sample reports that show what a completed transparency log or high-risk technical documentation file actually looks like. They reduce overwhelm by routing readers to the single most relevant obligations instead of the entire 200+ page Regulation.

Because they are use-case first, they convert better than generic compliance hubs: a recruiter worrying about CV-screening bias finds a page that speaks about recruitment metrics and evidence retention, not just recitals. Every page links to free tools that produce concrete outputs (disclosure text, evidence registers, literacy matrices) so teams can start building their compliance artifacts today.

Each linked page follows the same operational structure so you know exactly what to expect:

• Use-case matrix showing provider vs. deployer splits and how the system interacts with high-risk or transparency triggers.
• Evidence checklist listing concrete artifacts (e.g., interaction logs, watermarking tests, human oversight records, AI literacy training records) without promising certification.
• Common mistakes section based on real implementation pitfalls.
• Next step with a direct CTA into the most relevant free tool or sample report.

All content is built from official sources: the consolidated AI Act text, Article 50 guidelines and Code of Practice drafts, AI literacy Q&A, and the guidelines on the AI system definition and prohibited practices.

• Treating every chatbot as purely low-risk. If the same system performs high-risk tasks (e.g., recruitment screening or credit decisions), the stricter obligations apply to that function.
• Confusing provider and deployer roles. A company building and releasing a recruitment AI has provider obligations (technical documentation, conformity assessment). A company using an off-the-shelf tool has deployer obligations (following instructions, monitoring, reporting serious incidents).
• Skipping AI literacy measures. Article 4 requires sufficient literacy tailored to technical knowledge, context, and affected persons. There is no mandatory AI officer or formal exam, but evidence of tailored training is expected.
• Assuming “human review” automatically removes transparency obligations for marketing content. The Code of Practice drafts emphasise that review must be meaningful and editorial responsibility clear; otherwise disclosure is still required.
• Waiting for final Code of Practice or guidelines before starting. Voluntary codes and guidelines are helpful but not a substitute for mapping your use case against the current Regulation text now.
• Ignoring machine-readable marking for generative outputs. Article 50(2) requires providers to enable detection “as far as technically feasible” taking state of the art into account.

1. Identify whether your team acts as provider, deployer, or both for the specific AI system.
2. Map the use case against Annex III (high-risk) and Article 50 (transparency) using the official guidelines on AI system definition.
3. Run the relevant free tool: Article 50 Disclosure Generator for chatbots and generative content, or Evidence Scanner for high-risk systems.
4. Build the first evidence artifacts (interaction logs, training records, oversight procedures) and store them in a version-controlled register.
5. Schedule AI literacy sessions tailored to the roles interacting with the system.
6. Bookmark the specific sector page and revisit when the system or use case changes.

Ready to start? Choose your sector above and go straight to the most relevant free tool. For transparency-heavy uses (chatbots, marketing content) open the Article 50 Disclosure Generator. For recruitment, HR, education or credit uses open the Evidence Scanner. Both produce ready-to-use outputs that make your next internal review or audit far easier.

Which sector page should I use if my system spans multiple functions? Start with the highest-risk function. A recruitment chatbot that also answers general HR queries should begin on the recruitment page because the high-risk classification drives most obligations. Document the split clearly and apply the stricter rules where they overlap.

Are sector pages legal advice? No. They translate official EU sources into operational checklists, examples, and tool-supported workflows. They are not a substitute for your own legal review or advice from qualified counsel. Always verify against the latest consolidated text on eur-lex.europa.eu and the AI Act Service Desk.

Do sector pages replace the main guides? No. They are the fastest on-ramp for use-case teams. The main guides (High Risk Ai Systems, Transparency Obligations, Ai Literacy) provide deeper horizontal explanations. Use sector pages to route, then dive into the full guides for shared topics such as risk management or post-market monitoring.

Sources (all official)

• AI Act Regulation (EU) 2024/1689 (eur-lex.europa.eu)
• Article 50 Transparency FAQ and Code of Practice materials (digital-strategy.ec.europa.eu)
• AI Literacy Q&A and overview (digital-strategy.ec.europa.eu)
• Guidelines on AI system definition and prohibited practices (published February 2025)
• AI Act Service Desk timeline and Article 50 helper (ai-act-service-desk.ec.europa.eu)

All claims above are grounded in these primary sources. Proposed changes are explicitly labeled as such.