{"id":2246,"date":"2026-07-20T18:16:44","date_gmt":"2026-07-20T18:16:44","guid":{"rendered":"https:\/\/www.deepwraite.de\/eu-contributions\/"},"modified":"2026-09-04T14:17:16","modified_gmt":"2026-09-04T14:17:16","slug":"eu-contributions","status":"publish","type":"page","link":"https:\/\/www.deepwraite.de\/en\/eu-contributions\/","title":{"rendered":"EU contributions"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-page\" data-elementor-id=\"2246\" class=\"elementor elementor-2246 elementor-2242\" data-elementor-post-type=\"page\">\n\t\t\t\t<div class=\"elementor-element elementor-element-7f049027 e-flex e-con-boxed e-con e-parent\" data-id=\"7f049027\" data-element_type=\"container\" data-e-type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-26bccc63 elementor-widget elementor-widget-text-editor\" data-id=\"26bccc63\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t\t\t\t\t\t\n\n\n<style>\n  .dpw-eu{\n    --bg:#111110; --panel:rgba(255,255,255,.05); --ink:#ffffff;\n    --muted:#b4b4b4; --line:#FFFFFF0D; --accent:#0ee2cf;\n    --accent-soft:rgba(14,226,207,.14);\n    background:var(--bg); color:var(--ink); line-height:1.62;\n    font-family:\"Inter\",\"Helvetica Neue\",-apple-system,BlinkMacSystemFont,\"Segoe UI\",Roboto,Arial,sans-serif;\n    padding:8px 0 40px;\n  }\n  .dpw-eu *{box-sizing:border-box}\n  .dpw-eu .wrap{max-width:840px;margin:0 auto;padding:0 24px}\n\n  .dpw-eu .hero{position:relative;padding:56px 0 44px;overflow:hidden;border-bottom:1px solid var(--line)}\n  .dpw-eu .hero:before{content:\"\";position:absolute;right:-160px;top:-120px;width:420px;height:420px;\n    background:radial-gradient(circle at center,var(--accent-soft),transparent 62%);pointer-events:none}\n  .dpw-eu .kicker{position:relative;font-size:.78rem;letter-spacing:.18em;text-transform:uppercase;\n    color:var(--accent);font-weight:700;margin:0 0 16px}\n  .dpw-eu .title{position:relative;font-size:2.25rem;line-height:1.16;margin:0 0 18px;letter-spacing:-.015em;font-weight:700;color:var(--ink)}\n  .dpw-eu .lede{position:relative;font-size:1.16rem;color:var(--muted);margin:0;max-width:62ch}\n  .dpw-eu .tag{position:relative;margin-top:24px;font-size:.9rem;color:var(--muted)}\n  .dpw-eu .tag strong{color:var(--ink)}\n\n  .dpw-eu .contrib{padding:46px 0;border-bottom:1px solid var(--line)}\n  .dpw-eu .sec-title{font-size:1.55rem;margin:0 0 6px;letter-spacing:-.01em;font-weight:700;color:var(--ink)}\n  .dpw-eu .sec-title a{color:var(--ink);text-decoration:none}\n  .dpw-eu .meta{font-size:.84rem;color:var(--muted);margin:0 0 24px}\n  .dpw-eu .meta .dot{margin:0 9px;opacity:.5}\n  .dpw-eu .label{font-size:.78rem;letter-spacing:.13em;text-transform:uppercase;color:var(--accent);\n    margin:28px 0 12px;font-weight:700}\n  .dpw-eu p{margin:0 0 14px;color:var(--ink)}\n  .dpw-eu ul.points{list-style:none;margin:0;padding:0}\n  .dpw-eu ul.points li{position:relative;padding:14px 0 14px 26px;border-top:1px solid var(--line);color:var(--ink)}\n  .dpw-eu ul.points li:before{content:\"\";position:absolute;left:0;top:22px;width:9px;height:9px;\n    transform:rotate(45deg);background:var(--accent)}\n  .dpw-eu ul.points li b{color:var(--ink)}\n  .dpw-eu em{color:var(--ink);font-style:italic}\n  .dpw-eu .who{background:var(--panel);border:1px solid var(--line);border-radius:14px;\n    padding:16px 20px;margin-top:20px;font-size:.96rem;color:var(--muted)}\n  .dpw-eu .who span{color:var(--accent);font-weight:700}\n  .dpw-eu .foot{padding:42px 0 8px;color:var(--muted);font-size:.9rem}\n  .dpw-eu .foot a{color:var(--accent);text-decoration:none}\n  .dpw-eu .note{font-size:.82rem;border-left:2px solid var(--accent);padding:4px 0 4px 16px;margin-top:8px;color:var(--muted)}\n<\/style>\n\n<div class=\"dpw-eu\">\n  <div class=\"wrap\">\n\n    <div class=\"hero\">\n      <p class=\"kicker\">Deepwraite\u00ae \u00b7 EU rule-making<\/p>\n      <h2 class=\"title\">Our contributions to the EU AI Act guidelines<\/h2>\n      <p class=\"lede\">We contribute to how the AI Act is interpreted \u2013 from the perspective usually missing in these consultations: small and medium-sized enterprises, skilled trades and municipalities. Concrete cases from advisory practice instead of theory.<\/p>\n      <p class=\"tag\">Listed in the <strong>Stakeholders&#8217; Catalogue of the Apply AI Alliance<\/strong> of the European Commission.<\/p>\n    <\/div>\n\n    <div class=\"contrib\" id=\"artikel-6\">\n      <h3 class=\"sec-title\"><a href=\"#artikel-6\">Article 6 \u2013 Classification of high-risk AI systems<\/a><\/h3>\n      <p class=\"meta\">EU Commission targeted consultation on the draft guidelines under Article 6 AI Act (Annex III)<span class=\"dot\">\u2022<\/span>submitted on 20 July 2026<\/p>\n\n      <h4 class=\"label\">What it&#8217;s about<\/h4>\n      <p>The guidelines determine when an AI system counts as &#8220;high-risk&#8221; \u2013 and thereby triggers extensive obligations. For small businesses, a single question of interpretation often decides whether they are affected or not. In practice, classification is rarely a question of interpretation but of application.<\/p>\n\n      <h4 class=\"label\">What we contributed<\/h4>\n      <ul class=\"points\">\n        <li><b>The Article 6(3) filter, translated into SME reality.<\/b> The dividing line between a &#8220;narrow procedural task&#8221; and a substantive judgement is explained using school and visa examples \u2013 not transferable to invoice classification, email triage or quote review. We contributed in\/out examples from the skilled trades and SMEs.<\/li>\n        <li><b>Agentic and multi-agent systems.<\/b> When does a chained agent pipeline lose the filter? An operational test is missing \u2013 we provide a fully worked municipal example.<\/li>\n        <li><b>Proportionality of registration.<\/b> Even systems correctly classified as <em>not<\/em> high-risk must be documented and registered (Art. 6(4)). Counter-intuitive and little known for micro-enterprises \u2013 we ask for an SME-proportionate path.<\/li>\n        <li><b>Critical infrastructure \/ municipal utilities.<\/b> When does a predictive-maintenance model tip into a &#8220;safety component&#8221;? What role does an independent, non-AI safety layer play? In\/out examples from water and district-heating networks.<\/li>\n        <li><b>Employment \/ SME HR.<\/b> Recruiting filters (4a), the materiality threshold for shift planning and task allocation (4b), and a due-diligence checklist for using third-party applicant-tracking systems.<\/li>\n      <\/ul>\n\n      <div class=\"who\"><span>Who this is relevant for:<\/span> SMEs, skilled trades, municipal administrations and utilities, mid-sized employers \u2013 anyone using AI who needs to know whether they fall into &#8220;high-risk&#8221;.<\/div>\n    <\/div>\n\n    <div class=\"contrib\" id=\"artikel-50\">\n      <h3 class=\"sec-title\"><a href=\"#artikel-50\">Article 50 \u2013 Transparency obligations for AI systems<\/a><\/h3>\n      <p class=\"meta\">EU Commission targeted consultation on the draft guidelines under Article 50 AI Act<span class=\"dot\">\u2022<\/span>submitted on 1 June 2026<\/p>\n\n      <h4 class=\"label\">What it&#8217;s about<\/h4>\n      <p>Article 50 governs labelling and disclosure obligations: for AI interaction, AI-generated content, deepfakes and certain texts. In practice, it is often unclear to SMEs and municipalities which obligation applies to whom \u2013 and how it can be met technically.<\/p>\n\n      <h4 class=\"label\">What we contributed<\/h4>\n      <ul class=\"points\">\n        <li><b>Role allocation: provider vs. deployer.<\/b> Anyone deploying a commercial SaaS or LLM product under their own name remains the deployer \u2013 the upstream supplier is the provider. We address the gap when the provider cannot technically deliver the marking.<\/li>\n        <li><b>Watermarking vs. provenance.<\/b> The difference between embedded, model-side watermarks and after-the-fact provenance (C2PA) \u2013 and the GPAI downstream problem when marking is only &#8220;encouraged&#8221; rather than mandatory.<\/li>\n        <li><b>Transparency in interaction (50(1)).<\/b> Who discloses in multi-agent cascades? Retention for voice bots; consolidated disclosure across the DSA, UCPD and GDPR.<\/li>\n        <li><b>Biometrics and emotion (50(3)).<\/b> A negative list for standard municipal tools \u2013 e.g. plain speaker attribution when taking minutes, which is not categorisation.<\/li>\n        <li><b>Deepfakes and texts (50(4)).<\/b> Workable &#8220;editorial responsibility&#8221; for small municipalities, a feasible audience assessment, and role allocation in advertising-agency constellations.<\/li>\n        <li><b>SME proportionality.<\/b> The SME fine cap (Art. 99(6)) and the asymmetry between Code-of-Practice signatories and small businesses.<\/li>\n      <\/ul>\n\n      <div class=\"who\"><span>Who this is relevant for:<\/span> SMEs, skilled trades, municipal administrations and advertising agencies \u2013 anyone publishing AI interaction, AI-generated content or deepfakes.<\/div>\n    <\/div>\n\n    <div class=\"foot\">\n      <p>Both contributions are practitioner submissions from ongoing advisory mandates \u2013 not legal advice. Case examples anonymised.<\/p>\n      <p class=\"note\">Deepwraite\u00ae is the AI-governance practice of L&amp;L Kommunikationskonzepte GmbH. We enable SMEs, skilled trades and municipalities to adopt AI quickly, measurably and compliantly \u2013 in line with the EU AI Act and ISO\/IEC 42001. <em>Artificial intelligence that works. Governance that holds.<\/em><\/p>\n      <p style=\"margin-top:18px\"><a href=\"https:\/\/www.deepwraite.de\/en\/\">www.deepwraite.de<\/a><\/p>\n    <\/div>\n\n  <\/div>\n<\/div>\n\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>Deepwraite\u00ae \u00b7 EU rule-making Our contributions to the EU AI Act guidelines We contribute to how the AI Act is interpreted \u2013 from the perspective usually missing in these consultations: small and medium-sized enterprises, skilled trades and municipalities. Concrete cases from advisory practice instead of theory. Listed in the Stakeholders&#8217; Catalogue of the Apply AI [&hellip;]<\/p>\n","protected":false},"author":2,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-2246","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/www.deepwraite.de\/en\/wp-json\/wp\/v2\/pages\/2246","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.deepwraite.de\/en\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/www.deepwraite.de\/en\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/www.deepwraite.de\/en\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.deepwraite.de\/en\/wp-json\/wp\/v2\/comments?post=2246"}],"version-history":[{"count":9,"href":"https:\/\/www.deepwraite.de\/en\/wp-json\/wp\/v2\/pages\/2246\/revisions"}],"predecessor-version":[{"id":3248,"href":"https:\/\/www.deepwraite.de\/en\/wp-json\/wp\/v2\/pages\/2246\/revisions\/3248"}],"wp:attachment":[{"href":"https:\/\/www.deepwraite.de\/en\/wp-json\/wp\/v2\/media?parent=2246"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}