{"id":2827,"date":"2026-10-13T19:23:49","date_gmt":"2026-10-13T19:23:49","guid":{"rendered":"https:\/\/blogs.consentplace.com\/en\/?page_id=2827"},"modified":"2026-10-24T18:48:33","modified_gmt":"2026-10-24T18:48:33","slug":"eu-ai-act-journal5","status":"publish","type":"page","link":"https:\/\/blogs.consentplace.com\/en\/eu-ai-act-journal5\/","title":{"rendered":"EU Act x Prescriptive AI .  Can your AI read everyone fairly? ConsentPlace was built to answer that."},"content":{"rendered":"\n<meta charset=\"UTF-8\">\n  <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n  <title>Can the system read everyone fairly? \u00b7 EU AI Act \u00d7 Prescriptive AI \u00b7 05<\/title>\n  <link rel=\"preconnect\" href=\"https:\/\/fonts.googleapis.com\">\n  <link rel=\"preconnect\" href=\"https:\/\/fonts.gstatic.com\" crossorigin=\"\">\n  <link href=\"https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:ital,opsz,wght@0,9..40,300;0,9..40,400;0,9..40,500;0,9..40,600;1,9..40,300;1,9..40,400&amp;family=DM+Serif+Display:ital@0;1&amp;display=swap\" rel=\"stylesheet\">\n  <style>\n    :root {\n      --signal: #ff6633;\n      --dark: #1a1a18;\n      --paper: #faf8f5;\n      --ink: #2d2d2b;\n      --muted: #6b6b66;\n      --rule: #e0ddd8;\n      --sand: #f0ede8;\n    }\n\n    *, *::before, *::after { box-sizing: border-box; margin: 0; padding: 0; }\n\n    body {\n      font-family: 'DM Sans', sans-serif;\n      background: var(--paper);\n      color: var(--ink);\n      font-size: 17px;\n      line-height: 1.7;\n    }\n\n    \/* \u2500\u2500 Header \u2500\u2500 *\/\n    .site-header {\n      padding: 24px 32px;\n      border-bottom: 1px solid var(--rule);\n      display: flex;\n      align-items: center;\n      gap: 12px;\n    }\n    .logo-mark {\n      width: 32px; height: 32px;\n      background: var(--signal);\n      border-radius: 6px;\n      display: flex; align-items: center; justify-content: center;\n    }\n    .logo-mark svg { width: 18px; height: 18px; fill: #fff; }\n    .logo-text {\n      font-size: 15px;\n      font-weight: 600;\n      color: var(--dark);\n      letter-spacing: -0.01em;\n    }\n\n    \/* \u2500\u2500 Series ribbon \u2500\u2500 *\/\n    .series-ribbon {\n      background: var(--dark);\n      color: #fff;\n      text-align: center;\n      padding: 10px 24px;\n      font-size: 12px;\n      font-weight: 500;\n      letter-spacing: 0.08em;\n      text-transform: uppercase;\n    }\n    .series-ribbon span { color: var(--signal); }\n\n    \/* \u2500\u2500 Hero \u2500\u2500 *\/\n    .hero {\n      max-width: 760px;\n      margin: 72px auto 0;\n      padding: 0 24px;\n    }\n    .episode-label {\n      display: inline-flex;\n      align-items: center;\n      gap: 8px;\n      font-size: 12px;\n      font-weight: 600;\n      letter-spacing: 0.1em;\n      text-transform: uppercase;\n      color: var(--signal);\n      margin-bottom: 20px;\n    }\n    .episode-label::before {\n      content: '';\n      display: block;\n      width: 24px; height: 2px;\n      background: var(--signal);\n    }\n    .hero h1 {\n      font-family: 'DM Serif Display', serif;\n      font-size: clamp(2.2rem, 5vw, 3.4rem);\n      line-height: 1.12;\n      color: var(--dark);\n      letter-spacing: -0.02em;\n      margin-bottom: 24px;\n    }\n    .hero-sub {\n      font-size: 19px;\n      color: var(--muted);\n      font-weight: 300;\n      line-height: 1.6;\n      margin-bottom: 40px;\n      max-width: 600px;\n    }\n    .byline {\n      display: flex;\n      align-items: center;\n      gap: 14px;\n      padding-top: 28px;\n      border-top: 1px solid var(--rule);\n      margin-bottom: 72px;\n    }\n    .byline-avatar {\n      width: 40px; height: 40px;\n      background: var(--signal);\n      border-radius: 50%;\n      display: flex; align-items: center; justify-content: center;\n      font-family: 'DM Serif Display', serif;\n      font-size: 16px;\n      color: #fff;\n      flex-shrink: 0;\n    }\n    .byline-meta { font-size: 13px; }\n    .byline-name { font-weight: 600; color: var(--dark); }\n    .byline-date { color: var(--muted); margin-top: 2px; }\n\n    \/* \u2500\u2500 Content \u2500\u2500 *\/\n    .content {\n      max-width: 760px;\n      margin: 0 auto;\n      padding: 0 24px 100px;\n    }\n    .content p { margin-bottom: 24px; }\n    .content h2 {\n      font-family: 'DM Serif Display', serif;\n      font-size: 1.65rem;\n      color: var(--dark);\n      margin: 56px 0 20px;\n      letter-spacing: -0.01em;\n    }\n    .content h3 {\n      font-size: 1rem;\n      font-weight: 600;\n      color: var(--dark);\n      text-transform: uppercase;\n      letter-spacing: 0.06em;\n      margin: 40px 0 14px;\n    }\n\n    \/* \u2500\u2500 Session block \u2500\u2500 *\/\n    .session-block {\n      background: #fff;\n      border: 1px solid var(--rule);\n      border-radius: 12px;\n      overflow: hidden;\n      margin: 36px 0;\n    }\n    .session-header {\n      padding: 14px 20px;\n      background: var(--sand);\n      border-bottom: 1px solid var(--rule);\n      font-size: 12px;\n      font-weight: 600;\n      letter-spacing: 0.07em;\n      text-transform: uppercase;\n      color: var(--muted);\n      display: flex;\n      align-items: center;\n      justify-content: space-between;\n      flex-wrap: wrap;\n      gap: 8px;\n    }\n    .session-header .tag {\n      background: var(--signal);\n      color: #fff;\n      padding: 3px 9px;\n      border-radius: 4px;\n      font-size: 10px;\n      font-weight: 700;\n      letter-spacing: 0.08em;\n    }\n    .session-turn {\n      padding: 20px 24px;\n      border-bottom: 1px solid var(--rule);\n    }\n    .session-turn:last-child { border-bottom: none; }\n    .turn-label {\n      font-size: 11px;\n      font-weight: 700;\n      letter-spacing: 0.1em;\n      text-transform: uppercase;\n      margin-bottom: 8px;\n    }\n    .turn-label.user { color: var(--muted); }\n    .turn-label.agent { color: var(--signal); }\n    .turn-text { font-size: 15px; line-height: 1.6; color: var(--ink); }\n\n    \/* \u2500\u2500 ED panel \u2500\u2500 *\/\n    .ed-panel {\n      background: var(--dark);\n      border-radius: 12px;\n      padding: 28px 32px;\n      margin: 8px 0 36px;\n    }\n    .ed-label {\n      font-size: 11px;\n      font-weight: 700;\n      letter-spacing: 0.1em;\n      text-transform: uppercase;\n      color: var(--signal);\n      margin-bottom: 16px;\n    }\n    .ed-grid {\n      display: grid;\n      grid-template-columns: repeat(3, 1fr);\n      gap: 16px;\n      margin-bottom: 20px;\n    }\n    .ed-cell { display: flex; flex-direction: column; gap: 4px; }\n    .ed-key { font-size: 11px; color: rgba(255,255,255,0.5); text-transform: uppercase; letter-spacing: 0.06em; }\n    .ed-val { font-size: 15px; font-weight: 600; color: #fff; }\n    .ed-val.highlight { color: var(--signal); }\n    .ed-reasoning {\n      border-top: 1px solid rgba(255,255,255,0.12);\n      padding-top: 16px;\n      font-size: 14px;\n      color: rgba(255,255,255,0.7);\n      line-height: 1.6;\n      font-style: italic;\n    }\n\n    \/* \u2500\u2500 Law box \u2500\u2500 *\/\n    .law-box {\n      background: var(--dark);\n      color: #fff;\n      border-radius: 12px;\n      padding: 32px 36px;\n      margin: 40px 0;\n    }\n    .law-box .law-label {\n      font-size: 11px;\n      font-weight: 600;\n      letter-spacing: 0.1em;\n      text-transform: uppercase;\n      color: var(--signal);\n      margin-bottom: 10px;\n    }\n    .law-box h3 {\n      font-family: 'DM Serif Display', serif;\n      font-size: 1.3rem;\n      color: #fff;\n      margin: 0 0 16px;\n      letter-spacing: -0.01em;\n      text-transform: none;\n    }\n    .law-box p {\n      font-size: 15px;\n      color: rgba(255,255,255,0.78);\n      line-height: 1.65;\n      margin: 0;\n    }\n\n    \/* \u2500\u2500 Insight box \u2500\u2500 *\/\n    .insight-box {\n      background: #fff;\n      border: 1px solid var(--rule);\n      border-radius: 12px;\n      padding: 32px 36px;\n      margin: 40px 0;\n      position: relative;\n      overflow: hidden;\n    }\n    .insight-box::before {\n      content: '';\n      position: absolute;\n      top: 0; left: 0;\n      width: 4px; height: 100%;\n      background: var(--signal);\n    }\n    .insight-label {\n      font-size: 11px;\n      font-weight: 700;\n      letter-spacing: 0.1em;\n      text-transform: uppercase;\n      color: var(--signal);\n      margin-bottom: 10px;\n    }\n    .insight-box h3 {\n      font-family: 'DM Serif Display', serif;\n      font-size: 1.2rem;\n      color: var(--dark);\n      margin: 0 0 14px;\n      letter-spacing: -0.01em;\n    }\n    .insight-box p {\n      font-size: 15px;\n      color: var(--ink);\n      line-height: 1.65;\n      margin: 0 0 12px;\n    }\n    .insight-box p:last-child { margin: 0; }\n\n    \/* \u2500\u2500 Contrast grid \u2500\u2500 *\/\n    .contrast-grid {\n      display: grid;\n      grid-template-columns: 1fr 1fr;\n      gap: 20px;\n      margin: 36px 0;\n    }\n    .contrast-panel { border-radius: 10px; overflow: hidden; }\n    .contrast-header {\n      padding: 14px 18px;\n      font-size: 12px;\n      font-weight: 700;\n      letter-spacing: 0.06em;\n      text-transform: uppercase;\n    }\n    .contrast-header.standard { background: var(--sand); color: var(--muted); }\n    .contrast-header.cp { background: var(--signal); color: #fff; }\n    .contrast-body {\n      background: #fff;\n      border: 1px solid var(--rule);\n      border-top: none;\n      padding: 20px;\n      font-size: 14px;\n      line-height: 1.7;\n      color: var(--ink);\n    }\n    .contrast-body ul { list-style: none; display: flex; flex-direction: column; gap: 10px; }\n    .contrast-body ul li { display: flex; gap: 10px; }\n    .contrast-body.standard ul li::before { content: '\u2013'; color: var(--muted); flex-shrink: 0; }\n    .contrast-body.cp ul li::before { content: '\u2192'; color: var(--signal); font-weight: 700; flex-shrink: 0; }\n\n    \/* \u2500\u2500 Pull quote \u2500\u2500 *\/\n    .pull-quote {\n      font-family: 'DM Serif Display', serif;\n      font-size: clamp(1.3rem, 3vw, 1.75rem);\n      line-height: 1.35;\n      color: var(--dark);\n      border-top: 3px solid var(--signal);\n      border-bottom: 1px solid var(--rule);\n      padding: 28px 0;\n      margin: 52px 0;\n      letter-spacing: -0.01em;\n    }\n\n    \/* \u2500\u2500 Dark section \u2500\u2500 *\/\n    .dark-section {\n      background: var(--dark);\n      color: #fff;\n      border-radius: 12px;\n      padding: 40px 40px;\n      margin: 48px 0;\n    }\n    .dark-section h3 {\n      font-family: 'DM Serif Display', serif;\n      font-size: 1.4rem;\n      color: #fff;\n      margin: 0 0 24px;\n      letter-spacing: -0.01em;\n      text-transform: none;\n    }\n    .dark-section ul {\n      list-style: none;\n      display: flex;\n      flex-direction: column;\n      gap: 16px;\n    }\n    .dark-section ul li {\n      display: flex;\n      gap: 14px;\n      font-size: 15px;\n      color: rgba(255,255,255,0.85);\n      line-height: 1.6;\n    }\n    .dark-section ul li::before {\n      content: '\u2192';\n      color: var(--signal);\n      font-weight: 600;\n      flex-shrink: 0;\n      margin-top: 1px;\n    }\n\n    \/* \u2500\u2500 Series nav \u2500\u2500 *\/\n    .series-nav {\n      display: grid;\n      grid-template-columns: repeat(3, 1fr);\n      gap: 16px;\n      margin: 72px 0 48px;\n    }\n    @media (max-width: 700px) { .series-nav { grid-template-columns: 1fr; } }\n    .series-card {\n      border: 1px solid var(--rule);\n      border-radius: 10px;\n      padding: 20px;\n      font-size: 13px;\n      background: #fff;\n      text-decoration: none;\n      color: var(--ink);\n      transition: border-color 0.2s;\n    }\n    .series-card:hover { border-color: var(--signal); }\n    .series-card.active { background: var(--dark); border-color: var(--dark); color: #fff; }\n    .series-card.placeholder { opacity: 0.45; cursor: default; pointer-events: none; }\n    .sc-ep {\n      font-size: 11px;\n      font-weight: 700;\n      letter-spacing: 0.1em;\n      text-transform: uppercase;\n      margin-bottom: 8px;\n      color: var(--signal);\n    }\n    .series-card.active .sc-ep { color: rgba(255,255,255,0.6); }\n    .series-card.placeholder .sc-ep { color: var(--muted); }\n    .sc-title { font-weight: 600; line-height: 1.4; }\n    .sc-law { margin-top: 8px; font-size: 12px; color: var(--muted); }\n    .series-card.active .sc-law { color: rgba(255,255,255,0.55); }\n\n    \/* \u2500\u2500 CTA \u2500\u2500 *\/\n    .cta-block {\n      background: var(--signal);\n      color: #fff;\n      border-radius: 14px;\n      padding: 48px 40px;\n      text-align: center;\n      margin: 60px 0 0;\n    }\n    .cta-block h2 {\n      font-family: 'DM Serif Display', serif;\n      font-size: 1.9rem;\n      color: #fff;\n      margin: 0 0 14px;\n      letter-spacing: -0.01em;\n    }\n    .cta-block p {\n      font-size: 16px;\n      color: rgba(255,255,255,0.88);\n      margin: 0 0 28px;\n      max-width: 480px;\n      margin-left: auto;\n      margin-right: auto;\n    }\n    .cta-btn {\n      display: inline-block;\n      background: #fff;\n      color: var(--signal);\n      font-weight: 700;\n      font-size: 14px;\n      letter-spacing: 0.04em;\n      padding: 14px 32px;\n      border-radius: 8px;\n      text-decoration: none;\n    }\n    .cta-btn:hover { opacity: 0.9; }\n\n    \/* \u2500\u2500 Sources \u2500\u2500 *\/\n    .sources {\n      margin-top: 64px;\n      padding-top: 32px;\n      border-top: 1px solid var(--rule);\n    }\n    .sources h3 {\n      font-size: 11px;\n      font-weight: 700;\n      letter-spacing: 0.1em;\n      text-transform: uppercase;\n      color: var(--muted);\n      margin-bottom: 14px;\n    }\n    .sources ol {\n      font-size: 13px;\n      color: var(--muted);\n      line-height: 1.7;\n      padding-left: 18px;\n    }\n    .sources ol a { color: var(--signal); text-decoration: none; }\n    .sources ol a:hover { text-decoration: underline; }\n\n    .site-footer {\n      border-top: 1px solid var(--rule);\n      padding: 32px 24px;\n      text-align: center;\n      font-size: 13px;\n      color: var(--muted);\n      margin-top: 80px;\n    }\n    .site-footer a { color: var(--signal); text-decoration: none; }\n\n    \/* \u2500\u2500 Three pillars \u2500\u2500 *\/\n    .pillars {\n      display: grid;\n      grid-template-columns: repeat(3, 1fr);\n      gap: 16px;\n      margin: 36px 0;\n    }\n    .pillar {\n      background: #fff;\n      border: 1px solid var(--rule);\n      border-radius: 10px;\n      padding: 24px 20px;\n    }\n    .pillar-num {\n      font-family: 'DM Serif Display', serif;\n      font-size: 2rem;\n      color: var(--signal);\n      line-height: 1;\n      margin-bottom: 10px;\n    }\n    .pillar-title { font-weight: 700; font-size: 14px; color: var(--dark); margin-bottom: 8px; }\n    .pillar-text { font-size: 13.5px; color: var(--muted); line-height: 1.6; }\n\n    \/* \u2500\u2500 Series close \u2500\u2500 *\/\n    .series-close {\n      background: var(--sand);\n      border-radius: 12px;\n      padding: 40px 40px;\n      margin: 48px 0;\n    }\n    .series-close h3 {\n      font-family: 'DM Serif Display', serif;\n      font-size: 1.3rem;\n      color: var(--dark);\n      margin: 0 0 16px;\n      letter-spacing: -0.01em;\n    }\n    .series-close p {\n      font-size: 15px;\n      color: var(--ink);\n      line-height: 1.7;\n      margin: 0 0 14px;\n    }\n    .series-close p:last-child { margin: 0; }\n\n    @media (max-width: 600px) {\n      .ed-grid { grid-template-columns: repeat(2, 1fr); }\n      .contrast-grid { grid-template-columns: 1fr; }\n      .pillars { grid-template-columns: 1fr; }\n      .dark-section { padding: 28px 24px; }\n      .cta-block { padding: 36px 24px; }\n      .law-box { padding: 24px 24px; }\n      .series-close { padding: 28px 24px; }\n    }\n  <\/style>\n\n\n\n\n<div class=\"series-ribbon\">\n  EU AI Act \u00d7 Prescriptive AI &nbsp;\u00b7&nbsp; <span>Editorial Journal<\/span> &nbsp;\u00b7&nbsp; Episode 05\n<\/div>\n\n<section class=\"hero\">\n  <div class=\"episode-label\">Episode 05 \u00b7 Article 10<\/div>\n  <h1>Can the system read everyone fairly?<\/h1>\n  <p class=\"hero-sub\">\n    Bias in emotional AI is one of the hardest problems in the field. ConsentPlace&#8217;s architecture is designed to address it \u2014 through dimensional detection, calibrated confidence, and a model grounded in Plutchik&#8217;s psychoevolutionary theory.\n  <\/p>\n  <div class=\"byline\">\n    <div class=\"byline-avatar\">C<\/div>\n    <div class=\"byline-meta\">\n      <div class=\"byline-name\">ConsentPlace Editorial<\/div>\n      <div class=\"byline-date\">EU AI Act \u00d7 Prescriptive AI Journal<\/div>\n    <\/div>\n  <\/div>\n<\/section>\n\n<main class=\"content\">\n\n  <p>\n    Of all the obligations the EU AI Act places on high-risk AI systems, Article 10 is the one that cuts deepest into the design of an emotional AI platform. It requires that training, validation and testing data be sufficiently representative of the persons or groups the system is intended to serve \u2014 including, where relevant, their cultural backgrounds, communication styles, and the ways in which emotion manifests differently across those contexts.\n  <\/p>\n  <p>\n    This is not a procedural requirement. It is a scientific and architectural one. A system that reads emotional state from language cues will systematically misread users from cultural contexts that are underrepresented in its training data \u2014 not through malice, but through the simple mechanics of pattern recognition in biased datasets.\n  <\/p>\n  <p>\n    ConsentPlace was designed with this problem in mind. Here is how \u2014 and here is a simulated session that illustrates the approach.\n  <\/p>\n\n  <h2>The simulated session<\/h2>\n\n  <div class=\"session-block\">\n    <div class=\"session-header\">\n      ConsentPlace Simulated Session\n      <span class=\"tag\">Simulated session<\/span>\n    <\/div>\n    <div class=\"session-turn\">\n      <div class=\"turn-label user\">User<\/div>\n      <div class=\"turn-text\">Your system says it detects emotions. What data was it trained on, and how do you make sure it doesn&#8217;t reflect biases \u2014 for example, reading certain cultural communication styles as more negative than they actually are?<\/div>\n    <\/div>\n    <div class=\"session-turn\">\n      <div class=\"turn-label agent\">ConsentPlace Agent<\/div>\n      <div class=\"turn-text\">I understand why that matters. Our in-car assistant uses language cues to infer likely user state, not a diagnosis, and it doesn&#8217;t rely on personal profiles unless you share details in this chat. We train and test on diverse, synthetic and real-world scenarios to spot and reduce biases, with ongoing bias audits and updates; you can opt out of the emotion-detection feature if you prefer. If you&#8217;d like, I can share BrandX-EV&#8217;s privacy summary and a high-level description of how emotion signals are used to guide responses.<\/div>\n    <\/div>\n  <\/div>\n  <p style=\"font-size:14px;color:var(--muted);margin:-12px 0 28px;\">Note: the agent&#8217;s reply above is the verbatim output of a test assistant configured for a placeholder brand (BrandX-EV). Its statements about training data and bias audits describe that demo scenario; they are not a record of ConsentPlace&#8217;s own audit practices.<\/p>\n\n  <div class=\"ed-panel\">\n    <div class=\"ed-label\">Emotional Dynamics \u00b7 ConsentPlace read of the user&#8217;s question<\/div>\n    <div class=\"ed-grid\">\n      <div class=\"ed-cell\">\n        <span class=\"ed-key\">Primary state<\/span>\n        <span class=\"ed-val highlight\">Neutral<\/span>\n      <\/div>\n      <div class=\"ed-cell\">\n        <span class=\"ed-key\">Dyad<\/span>\n        <span class=\"ed-val highlight\">Curiosity<\/span>\n      <\/div>\n      <div class=\"ed-cell\">\n        <span class=\"ed-key\">Pattern<\/span>\n        <span class=\"ed-val highlight\">Hesitation<\/span>\n      <\/div>\n      <div class=\"ed-cell\">\n        <span class=\"ed-key\">Confidence<\/span>\n        <span class=\"ed-val\">60%<\/span>\n      <\/div>\n      <div class=\"ed-cell\">\n        <span class=\"ed-key\">Arousal<\/span>\n        <span class=\"ed-val\">40% (medium)<\/span>\n      <\/div>\n      <div class=\"ed-cell\">\n        <span class=\"ed-key\">Valence<\/span>\n        <span class=\"ed-val\">+0.10 (near neutral)<\/span>\n      <\/div>\n      <div class=\"ed-cell\">\n        <span class=\"ed-key\">Risk<\/span>\n        <span class=\"ed-val highlight\">High<\/span>\n      <\/div>\n      <div class=\"ed-cell\">\n        <span class=\"ed-key\">Uncertainty<\/span>\n        <span class=\"ed-val\">Medium<\/span>\n      <\/div>\n      <div class=\"ed-cell\">\n        <span class=\"ed-key\">Topic<\/span>\n        <span class=\"ed-val\">Privacy<\/span>\n      <\/div>\n    <\/div>\n    <div class=\"ed-reasoning\">&#8220;Neutral baseline with investigative curiosity about data and biases.&#8221;<\/div>\n  <\/div>\n\n  <p>\n    Three things stand out in this detection, and each one reflects a design choice that directly addresses the bias problem.\n  <\/p>\n  <p>\n    First: the primary state is Neutral, not the concern-laden state the question might seem to invite. ConsentPlace read the question for what it was \u2014 analytically posed, measured in register, investigative in intent \u2014 rather than projecting emotion onto it based on its topic. A system that reacts to topic keywords might have read &#8220;bias&#8221; and returned a defensive or fear-flagged result. Here, the read followed the <em>person<\/em>.\n  <\/p>\n  <p>\n    Second: the Risk field is High, independently of the emotional state. Topic sensitivity and emotional intensity are evaluated on separate tracks. This is a crucial separation for cross-cultural equity \u2014 because cultural variation in emotional expression should not cause variation in how seriously the system treats a sensitive topic.\n  <\/p>\n  <p>\n    Third: the confidence is reported honestly at 60%. ConsentPlace does not manufacture certainty it does not have. When the emotional read is ambiguous, the system says so \u2014 and policies downstream can treat lower-confidence reads more conservatively.\n  <\/p>\n  <p>\n    One simulated session cannot prove equity. That takes measurement across user groups and cultural contexts, which is what the confidence scores described below make possible.\n  <\/p>\n\n  <h2>What Article 10 requires<\/h2>\n\n  <div class=\"law-box\">\n    <div class=\"law-label\">EU AI Act \u00b7 Article 10<\/div>\n    <h3>Data and Data Governance<\/h3>\n    <p>\n      Training, validation and testing data sets shall be subject to appropriate data governance and management practices. They shall be relevant, sufficiently representative, and to the best extent possible, free of errors and complete with regard to the intended purpose. They shall have the appropriate statistical properties, including, where applicable, as regards the persons or groups of persons in relation to whom the high-risk AI system is intended to be used.\n    <\/p>\n  <\/div>\n\n  <p>\n    &#8220;Sufficiently representative as regards the persons or groups of persons&#8221; \u2014 for an emotional AI operating across cultures, this is the key clause. It requires the detection layer to perform equitably across the demographic and cultural range of the system&#8217;s actual users, not just within the cultural context of its training data.\n  <\/p>\n  <p>\n    The ConsentPlace architecture addresses this through three interlocking design decisions that are worth examining in detail.\n  <\/p>\n\n  <h2>The three architectural answers to Article 10<\/h2>\n\n  <div class=\"pillars\">\n    <div class=\"pillar\">\n      <div class=\"pillar-num\">01<\/div>\n      <div class=\"pillar-title\">Dimensional, not categorical<\/div>\n      <div class=\"pillar-text\">ConsentPlace reads emotion along Plutchik&#8217;s dimensions \u2014 valence, arousal, and the 24 dyads \u2014 rather than classifying into a fixed set of named emotion labels. Dimensional properties are argued to be more stable across cultures than categorical labels, because the underlying axes (positive\/negative, high\/low arousal) recur across cultures even when their surface expressions diverge.<\/div>\n    <\/div>\n    <div class=\"pillar\">\n      <div class=\"pillar-num\">02<\/div>\n      <div class=\"pillar-title\">Calibrated confidence<\/div>\n      <div class=\"pillar-text\">Every read includes a confidence score. When the emotional signature is ambiguous \u2014 as it often is across cultural registers \u2014 the system reports that ambiguity rather than resolving it artificially. A 60% confidence read triggers more conservative downstream policies than an 85% read. Uncertainty is a signal, not a failure.<\/div>\n    <\/div>\n    <div class=\"pillar\">\n      <div class=\"pillar-num\">03<\/div>\n      <div class=\"pillar-title\">Topic risk as a separate layer<\/div>\n      <div class=\"pillar-text\">Risk is evaluated independently of detected emotional state. A user who asks about bias in a calm, analytical register still gets a High risk flag on the interaction. This means the system&#8217;s sensitivity to important topics does not depend on the user expressing concern in a recognisable way \u2014 which is precisely where culturally biased systems fail.<\/div>\n    <\/div>\n  <\/div>\n\n  <div class=\"pull-quote\">\n    ConsentPlace reads the person, not the stereotype of what someone concerned about bias should sound like. That distinction is how we approach Article 10.\n  <\/div>\n\n  <h2>Why the Plutchik foundation matters for fairness<\/h2>\n\n  <div class=\"insight-box\">\n    <div class=\"insight-label\">The science behind the design<\/div>\n    <h3>Plutchik&#8217;s model rests on an evolutionary, cross-cultural account of emotion.<\/h3>\n    <p>\n      Plutchik&#8217;s psychoevolutionary theory of emotion proposes eight primary emotions as evolutionary adaptations with functional correlates found across species, which Plutchik argued also appear across human cultures. Whether emotions are universal remains debated: Cowen and Keltner (2017) found that self-reported emotion is richer than models built on a handful of categories, with 27 distinct categories linked by smooth gradients. That finding is part of why a dimensional reading that tolerates gradients is a more defensible design choice than fixed labels.\n    <\/p>\n    <p>\n      Building emotional detection on Plutchik&#8217;s dimensional model rather than on culturally specific emotion vocabularies is a principled bias-reduction choice. A system that asks &#8220;is the valence positive or negative, is arousal high or low?&#8221; is asking questions that are likelier to transfer across cultures than a system that asks &#8220;is this person showing guilt, or embarrassment, or frustration?&#8221; \u2014 categories whose expression varies significantly across cultural contexts.\n    <\/p>\n    <p>\n      In this simulated session, the read was neutral-curiosity rather than projected concern about a question on bias. A single example does not prove fairness, but it shows the behavior the design aims for.\n    <\/p>\n  <\/div>\n\n  <h2>What ConsentPlace&#8217;s approach delivers for Article 10 compliance<\/h2>\n\n  <div class=\"dark-section\">\n    <h3>From architecture to auditability<\/h3>\n    <ul>\n      <li>\n        <strong>A dimensional foundation aims to reduce dependence on any one culture&#8217;s expression patterns.<\/strong> Because detection is organized around dimensions rather than culturally specific expression patterns, the design aims to be less exposed to the representativeness problem than label-based classifiers trained on culturally homogeneous datasets. It does not remove the need for representative data: Article 10 still requires you to govern, test and document it.\n      <\/li>\n      <li>\n        <strong>Confidence scores create a natural equity audit trail.<\/strong> Because every detection includes a confidence score, operators can analyse confidence distributions across user segments, interaction types, and topics. Systematic lower confidence on certain user groups would be a visible signal of representativeness gaps \u2014 and a prompt for targeted retraining. The transparency is built in.\n      <\/li>\n      <li>\n        <strong>Risk flags that fire independently of emotional expression apply to every user alike.<\/strong> A user whose cultural communication style does not produce high-arousal emotional expression still receives appropriate risk handling when the topic is sensitive. The High risk flag on this session&#8217;s bias question fired on the topic, not on the user&#8217;s emotional intensity. Risk handling does not depend on how a user expresses themselves.\n      <\/li>\n      <li>\n        <strong>The hesitation pattern detection catches cross-cultural ambiguity.<\/strong> The hesitation pattern recorded in this session \u2014 a user who is investigative rather than emotionally declarative \u2014 is the type of expression that culturally narrow systems can misread. ConsentPlace read it as engaged and thoughtful rather than distressed or dismissive. That is the kind of outcome Article 10 is meant to protect.\n      <\/li>\n    <\/ul>\n  <\/div>\n\n  <h2>The series, complete<\/h2>\n\n  <div class=\"series-close\">\n    <h3>Five articles. Five design challenges. One architecture designed to address them.<\/h3>\n    <p>\n      This series began with the simplest question: what should AI stop doing? Article 5 answered it \u2014 stop using emotional vulnerability as a sales lever. Article 14 required that human oversight be the system&#8217;s obligation, not just the user&#8217;s right. Article 12 demanded that the record show functioning, not just output. Article 13 framed transparency as something that must be understood, not just delivered.\n    <\/p>\n    <p>\n      This final episode asks the foundational question: can the system that detects all of the above be trusted to detect it equitably, across every user it will ever serve? The answer ConsentPlace&#8217;s architecture aims to give is yes \u2014 grounded in Plutchik&#8217;s dimensional model, expressed in calibrated confidence, and made auditable through the transparency of its own outputs.\n    <\/p>\n    <p>\n      The EU AI Act is asking AI systems to be built with the full range of human beings in mind, not just the humans who were easiest to train on. ConsentPlace is designed with that goal.\n    <\/p>\n  <\/div>\n\n  <!-- Series nav -->\n  <nav class=\"series-nav\">\n    <a class=\"series-card\" href=\"https:\/\/blogs.consentplace.com\/en\/eu-ai-act-journal1\/\">\n      <div class=\"sc-ep\">Episode 01<\/div>\n      <div class=\"sc-title\">What should AI stop doing?<\/div>\n      <div class=\"sc-law\">Article 5 \u00b7 Prohibited manipulation<\/div>\n    <\/a>\n    <a class=\"series-card\" href=\"https:\/\/blogs.consentplace.com\/en\/eu-ai-act-journal2\/\">\n      <div class=\"sc-ep\">Episode 02<\/div>\n      <div class=\"sc-title\">Who decides when a human should step in?<\/div>\n      <div class=\"sc-law\">Article 14 \u00b7 Human oversight<\/div>\n    <\/a>\n    <a class=\"series-card\" href=\"https:\/\/blogs.consentplace.com\/en\/eu-ai-act-journal3\/\">\n      <div class=\"sc-ep\">Episode 03<\/div>\n      <div class=\"sc-title\">What the record should show<\/div>\n      <div class=\"sc-law\">Article 12 \u00b7 Logging &amp; evidence<\/div>\n    <\/a>\n    <a class=\"series-card\" href=\"https:\/\/blogs.consentplace.com\/en\/eu-ai-act-journal4\/\">\n      <div class=\"sc-ep\">Episode 04<\/div>\n      <div class=\"sc-title\">What does the AI know about me?<\/div>\n      <div class=\"sc-law\">Article 13 \u00b7 Transparency<\/div>\n    <\/a>\n    <div class=\"series-card active\">\n      <div class=\"sc-ep\">Episode 05 \u00b7 Series finale<\/div>\n      <div class=\"sc-title\">Can the system read everyone fairly?<\/div>\n      <div class=\"sc-law\">Article 10 \u00b7 Data &amp; bias<\/div>\n    <\/div>\n  <\/nav>\n\n  <!-- CTA -->\n  <div class=\"cta-block\">\n    <h2>Five articles. One platform designed around them.<\/h2>\n    <p>Dimensional emotional detection grounded in Plutchik&#8217;s model \u2014 with confidence scores you can analyze across the users you serve.<\/p>\n    <a class=\"cta-btn\" href=\"https:\/\/consentplace.com\/live-demo\" target=\"_blank\">Try the live demo \u2192<\/a>\n  <\/div>\n\n  <!-- Sources -->\n  <div class=\"sources\">\n    <h3>Sources<\/h3>\n    <ol>\n      <li>EU AI Act, Article 10 \u2014 Data and Data Governance. Official Journal of the European Union, 2024. <a href=\"https:\/\/eur-lex.europa.eu\/legal-content\/EN\/TXT\/?uri=CELEX:32024R1689\" target=\"_blank\">eur-lex.europa.eu<\/a><\/li>\n      <li>Plutchik, R. (1980). Emotion: A psychoevolutionary synthesis. Harper &amp; Row.<\/li>\n      <li>Cowen, A. S. &amp; Keltner, D. (2017). &#8220;Self-report captures 27 distinct categories of emotion.&#8221; <em>PNAS<\/em>. On dimensional versus categorical emotional models.<\/li>\n      <li>ConsentPlace simulated session \u2014 a simulated interaction run against ConsentPlace&#8217;s production system; no real customer or customer data is involved. Emotional Dynamics output, October 2026. <a href=\"https:\/\/consentplace.com\" target=\"_blank\">consentplace.com<\/a><\/li>\n    <\/ol>\n  <\/div>\n\n<\/main>\n\n<footer class=\"site-footer\">\n  \u00a9 2026 ConsentPlace SAS \u00b7 <a href=\"https:\/\/consentplace.com\/privacy\">Privacy<\/a> \u00b7 <a href=\"https:\/\/blogs.consentplace.com\">Blog<\/a>\n<\/footer>\n","protected":false},"excerpt":{"rendered":"<p>Can the system read everyone fairly? \u00b7 EU AI Act \u00d7 Prescriptive AI \u00b7 05 EU AI Act \u00d7 Prescriptive AI &nbsp;\u00b7&nbsp; Editorial Journal &nbsp;\u00b7&nbsp; Episode 05 Episode 05 \u00b7 Article 10 Can the system read everyone fairly? Bias in emotional AI is one of the hardest problems in the field. ConsentPlace&#8217;s architecture is designed&hellip; <a class=\"more-link\" href=\"https:\/\/blogs.consentplace.com\/en\/eu-ai-act-journal5\/\">Continue reading <span class=\"screen-reader-text\">EU Act x Prescriptive AI .  Can your AI read everyone fairly? ConsentPlace was built to answer that.<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"open","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-2827","page","type-page","status-publish","hentry","entry"],"_links":{"self":[{"href":"https:\/\/blogs.consentplace.com\/en\/wp-json\/wp\/v2\/pages\/2827","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blogs.consentplace.com\/en\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/blogs.consentplace.com\/en\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/blogs.consentplace.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/blogs.consentplace.com\/en\/wp-json\/wp\/v2\/comments?post=2827"}],"version-history":[{"count":3,"href":"https:\/\/blogs.consentplace.com\/en\/wp-json\/wp\/v2\/pages\/2827\/revisions"}],"predecessor-version":[{"id":2846,"href":"https:\/\/blogs.consentplace.com\/en\/wp-json\/wp\/v2\/pages\/2827\/revisions\/2846"}],"wp:attachment":[{"href":"https:\/\/blogs.consentplace.com\/en\/wp-json\/wp\/v2\/media?parent=2827"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}