{"id":8662,"date":"2025-09-11T11:55:50","date_gmt":"2025-09-11T03:55:50","guid":{"rendered":"https:\/\/regask.com\/?p=8662"},"modified":"2025-10-15T14:47:31","modified_gmt":"2025-10-15T06:47:31","slug":"ki-gmp-dekodierung-neuer-bilder-anhang-22-regulierungsteams","status":"publish","type":"post","link":"https:\/\/regask.com\/de\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/","title":{"rendered":"KI in GMP: Entschl\u00fcsselung des neuen PIC\/S-Anhangs 22 f\u00fcr Regulierungsbeh\u00f6rden\u2026"},"content":{"rendered":"<p><span data-contrast=\"auto\">An\u00a0<\/span><b><span data-contrast=\"auto\">7. Juli 2025<\/span><\/b><span data-contrast=\"auto\">, Die\u00a0<\/span><a href=\"https:\/\/picscheme.org\/en\/picscheme\"><b><span data-contrast=\"none\">Kooperationsprogramm f\u00fcr Arzneimittelinspektionen (PIC\/S)<\/span><\/b><\/a><span data-contrast=\"auto\">\u00a0ver\u00f6ffentlicht\u00a0<\/span><b><span data-contrast=\"auto\">Anhang 22<\/span><\/b><span data-contrast=\"auto\">\u00a0zum PIC\/S GMP-Leitfaden. Dieser neue Anhang enth\u00e4lt detaillierte regulatorische Erwartungen f\u00fcr die Verwendung von\u00a0<\/span><b><span data-contrast=\"auto\">K\u00fcnstliche Intelligenz (KI) und maschinelles Lernen (ML)<\/span><\/b><span data-contrast=\"auto\">\u00a0In\u00a0<\/span><b><span data-contrast=\"auto\">GMP-regulierte Umgebungen<\/span><\/b><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Das Update schlie\u00dft L\u00fccken in\u00a0<\/span><b><span data-contrast=\"auto\">Anhang 11<\/span><\/b><span data-contrast=\"auto\">, die die sich entwickelnden KI\/ML-Technologien nicht vollst\u00e4ndig abdeckt. Ziel ist es, sicherzustellen\u00a0<\/span><b><span data-contrast=\"auto\">sichere, zuverl\u00e4ssige und transparente KI-Integration<\/span><\/b><span data-contrast=\"auto\">\u00a0in pharmazeutischen Betrieben, insbesondere in Bereichen, die direkt\u00a0<\/span><b><span data-contrast=\"auto\">Patientensicherheit, Produktqualit\u00e4t und Datenintegrit\u00e4t<\/span><\/b><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n\t<div id=\"gap-1885142268\" class=\"gap-element clearfix\" style=\"display:block; height:auto;\">\n\t\t\n<style>\n#gap-1885142268 {\n  padding-top: 30px;\n}\n<\/style>\n\t<\/div>\n\t\n<div class=\"row\"  id=\"row-202449557\">\n\n\t<div id=\"col-1281016918\" class=\"col small-12 large-12\"  >\n\t\t\t\t<div class=\"col-inner\"  >\n\t\t\t\n\t\t\t\n<h3><span class=\"TextRun MacChromeBold SCXW123257412 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW123257412 BCX0\" data-ccp-parastyle=\"heading 2\">Wesentliche Bestimmungen des Anhangs 22<\/span><\/span><span class=\"EOP SCXW123257412 BCX0\" data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:299,&quot;335559739&quot;:299}\">\u00a0<\/span><\/h3>\n\t<div id=\"gap-1495148918\" class=\"gap-element clearfix\" style=\"display:block; height:auto;\">\n\t\t\n<style>\n#gap-1495148918 {\n  padding-top: 10px;\n}\n<\/style>\n\t<\/div>\n\t\n<p><span data-contrast=\"auto\">Anhang 22 ist ein\u00a0<\/span><b><span data-contrast=\"auto\">Neuzugang<\/span><\/b><span data-contrast=\"auto\">\u00a0Es handelt sich nicht um eine \u00dcberarbeitung, sondern um eine Erg\u00e4nzung von Anhang 11 um KI-spezifische Anforderungen.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Der Anhang schr\u00e4nkt den Umfang der KI-Nutzung in GMP streng ein:\u00a0<\/span><b><span data-contrast=\"auto\">nur statische KI\/ML-Modelle mit deterministischen Ausgaben<\/span><\/b><span data-contrast=\"auto\">\u00a0k\u00f6nnen in kritischen Anwendungen eingesetzt werden. Systeme mit\u00a0<\/span><b><span data-contrast=\"auto\">dynamische Modelle, probabilistische Ergebnisse oder adaptives Lernen<\/span><\/b><span data-contrast=\"auto\">\u00a0Sind\u00a0<\/span><b><span data-contrast=\"auto\">in kritischen GMP-Kontexten verboten<\/span><\/b><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">Generative KI und gro\u00dfe Sprachmodelle (LLMs)<\/span><\/b><span data-contrast=\"auto\">\u00a0sind in kritischen Operationen nicht zul\u00e4ssig. In nicht-kritischen Kontexten ist ihre Verwendung nur akzeptabel mit\u00a0<\/span><b><span data-contrast=\"auto\">Human-in-the-Loop (HITL)-Aufsicht<\/span><\/b><span data-contrast=\"auto\">\u00a0durch qualifiziertes Personal.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Die Umsetzung muss\u00a0<\/span><b><span data-contrast=\"auto\">multidisziplin\u00e4r<\/span><\/b><span data-contrast=\"auto\">, einschlie\u00dflich\u00a0<\/span><b><span data-contrast=\"auto\">Qualit\u00e4tssicherung (QA), Datenwissenschaftler, IT-Teams, KMU und Berater<\/span><\/b><span data-contrast=\"auto\">. Jede Phase des KI-Lebenszyklus \u2013 von\u00a0<\/span><b><span data-contrast=\"auto\">Algorithmusauswahl und Training bis hin zu Validierung, Test und Bereitstellung<\/span><\/b><span data-contrast=\"auto\">\u2013 muss sein\u00a0<\/span><b><span data-contrast=\"auto\">risikoinformiert, nachvollziehbar und dokumentiert<\/span><\/b><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Die regulatorische Absicht ist klar: unvorhersehbares KI-Verhalten verhindern, Erkl\u00e4rbarkeit und \u00dcberpr\u00fcfbarkeit st\u00e4rken, bewahren\u00a0<\/span><b><span data-contrast=\"auto\">Datenintegrit\u00e4t<\/span><\/b><span data-contrast=\"auto\">und stellen Sie sicher\u00a0<\/span><b><span data-contrast=\"auto\">menschliche Verantwortung<\/span><\/b><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\n\t\n\n\t<div id=\"col-201183551\" class=\"col small-12 large-12\"  >\n\t\t\t\t<div class=\"col-inner\"  >\n\t\t\t\n\t\t\t\n<h3><span class=\"TextRun MacChromeBold SCXW72001948 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW72001948 BCX0\" data-ccp-parastyle=\"heading 2\">Warum es wichtig ist<\/span><\/span><span class=\"EOP SCXW72001948 BCX0\" data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:299,&quot;335559739&quot;:299}\">\u00a0<\/span><\/h3>\n\t<div id=\"gap-1068352077\" class=\"gap-element clearfix\" style=\"display:block; height:auto;\">\n\t\t\n<style>\n#gap-1068352077 {\n  padding-top: 10px;\n}\n<\/style>\n\t<\/div>\n\t\n<p><span class=\"TextRun SCXW218463272 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW218463272 BCX0\">Annex 22 erh\u00f6ht die Compliance-Erwartungen f\u00fcr Pharmahersteller, die KI in Betracht ziehen, erheblich. Unternehmen, die\u00a0<\/span><\/span><span class=\"TextRun MacChromeBold SCXW218463272 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW218463272 BCX0\">dynamische oder adaptive KI-Modelle<\/span><\/span><span class=\"TextRun SCXW218463272 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW218463272 BCX0\">\u00a0m\u00fcssen m\u00f6glicherweise\u00a0<\/span><\/span><span class=\"TextRun MacChromeBold SCXW218463272 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW218463272 BCX0\">Systemdesign und Strategie \u00fcberdenken<\/span><\/span><span class=\"TextRun SCXW218463272 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW218463272 BCX0\">Anforderungen f\u00fcr\u00a0<\/span><\/span><span class=\"TextRun MacChromeBold SCXW218463272 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW218463272 BCX0\">Erkl\u00e4rbarkeit, Testdatenunabh\u00e4ngigkeit und R\u00fcckverfolgbarkeit<\/span><\/span><span class=\"TextRun SCXW218463272 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW218463272 BCX0\">\u00a0k\u00f6nnte auch zunehmen\u00a0<\/span><\/span><span class=\"TextRun MacChromeBold SCXW218463272 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW218463272 BCX0\">Anforderungen an die Auditbereitschaft<\/span><\/span><span class=\"TextRun SCXW218463272 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW218463272 BCX0\">\u00a0Und\u00a0<\/span><span class=\"NormalTextRun SCXW218463272 BCX0\">erfordern<\/span><span class=\"NormalTextRun SCXW218463272 BCX0\">\u00a0engere funktions\u00fcbergreifende Zusammenarbeit.<\/span><\/span><span class=\"EOP SCXW218463272 BCX0\" data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\n\t\n\n\t<div id=\"col-1282911496\" class=\"col small-12 large-12\"  >\n\t\t\t\t<div class=\"col-inner\"  >\n\t\t\t\n\t\t\t\n<h3><span class=\"TextRun MacChromeBold SCXW217261047 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW217261047 BCX0\" data-ccp-parastyle=\"heading 2\">Relevanz<\/span><\/span><span class=\"EOP SCXW217261047 BCX0\" data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:299,&quot;335559739&quot;:299}\">\u00a0<\/span><\/h3>\n\t<div id=\"gap-395149001\" class=\"gap-element clearfix\" style=\"display:block; height:auto;\">\n\t\t\n<style>\n#gap-395149001 {\n  padding-top: 10px;\n}\n<\/style>\n\t<\/div>\n\t\n<p><span data-contrast=\"auto\">Dieses Update ist wichtig f\u00fcr:<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<ul>\n<li><b><span data-contrast=\"auto\">Regulierungs- und Qualit\u00e4tsteams<\/span><\/b><span data-contrast=\"auto\">\u00a0Sicherstellung der Einhaltung der GMP-Erwartungen<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"auto\">Data Scientists und IT-Spezialisten<\/span><\/b><span data-contrast=\"auto\">\u00a0Verwaltung der Entwicklung und Integration von KI-Systemen<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"auto\">Pharmahersteller<\/span><\/b><span data-contrast=\"auto\">\u00a0Planung oder Durchf\u00fchrung KI-gesteuerter Prozesse in GMP-regulierten Umgebungen<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/li>\n<\/ul>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\n\t\n\n\t<div id=\"col-968414504\" class=\"col small-12 large-12\"  >\n\t\t\t\t<div class=\"col-inner\"  >\n\t\t\t\n\t\t\t\n<h3><span class=\"TextRun MacChromeBold SCXW116595951 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW116595951 BCX0\" data-ccp-parastyle=\"heading 2\">N\u00e4chste Schritte f\u00fcr die Industrie<\/span><\/span><span class=\"EOP SCXW116595951 BCX0\" data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;134245418&quot;:true,&quot;134245529&quot;:true,&quot;335559738&quot;:299,&quot;335559739&quot;:299}\">\u00a0<\/span><\/h3>\n\t<div id=\"gap-1872446921\" class=\"gap-element clearfix\" style=\"display:block; height:auto;\">\n\t\t\n<style>\n#gap-1872446921 {\n  padding-top: 10px;\n}\n<\/style>\n\t<\/div>\n\t\n<p><span data-contrast=\"auto\">Organisationen sollten beginnen\u00a0<\/span><b><span data-contrast=\"auto\">interne Gap-Assessments<\/span><\/b><span data-contrast=\"auto\">\u00a0um aktuelle KI\/ML-Systeme mit den Anforderungen des Anhangs 22 zu vergleichen. Besonderes Augenmerk sollte auf\u00a0<\/span><b><span data-contrast=\"auto\">Modellvalidierung, Testgenauigkeit und Datentransparenz<\/span><\/b><span data-contrast=\"auto\">. Anhang 22 ist zwar neu, doch eine fr\u00fchzeitige Anpassung kann Compliance-St\u00f6rungen abmildern, sobald die Umsetzung verpflichtend wird.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Die Beteiligten werden aufgefordert, den Anhang sorgf\u00e4ltig zu pr\u00fcfen und sich darauf vorzubereiten, bei zuk\u00fcnftigen\u00a0<\/span><b><span data-contrast=\"auto\">Konsultationsphasen<\/span><\/b><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Beim \u00dcbergang von der Vorbereitung zur proaktiven Anpassung k\u00f6nnen Unternehmen von strukturierter Unterst\u00fctzung profitieren.\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">In diesem Fall,\u00a0<\/span><\/b><a href=\"https:\/\/regask.com\/de\/\"><b><span data-contrast=\"none\">RegASK<\/span><\/b><\/a><b><span data-contrast=\"none\">\u00a0ist eine KI-gesteuerte L\u00f6sung f\u00fcr Regulatory Intelligence und Workflow-Orchestrierung\u00a0<\/span><\/b><span data-contrast=\"none\">Wir helfen Unternehmen aus den Bereichen Biowissenschaften und Konsumg\u00fcter dabei, mit den sich \u00e4ndernden globalen Vorschriften und Richtlinien Schritt zu halten.\u00a0<\/span><b><span data-contrast=\"none\">Es kombiniert intelligente Dokumentation, automatisierte Workflows und integrierte Experten<\/span><\/b><span data-contrast=\"none\"> um eine schnellere Anpassung zu erm\u00f6glichen, was letztendlich das Betriebsrisiko reduziert.<\/span><\/p>\n<a href=\"https:\/\/regask.com\/de\/demo-buchen\/\" target=\"_blank\" class=\"button secondary\" rel=\"noopener\" style=\"border-radius:99px;\">\n\t\t<span>Demo buchen<\/span>\n\t<\/a>\n\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\n\t\n\n\t<div id=\"col-1284683210\" class=\"col small-12 large-12\"  >\n\t\t\t\t<div class=\"col-inner\"  >\n\t\t\t\n\t\t\t\n<div class=\"is-divider divider clearfix\" style=\"margin-top:2em;margin-bottom:2em;max-width:20%;height:1px;background-color:#a5acd9;\"><\/div>\n<h3><span class=\"TextRun MacChromeBold SCXW156847865 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW156847865 BCX0\">Definitionen und Akronyme:<\/span><\/span><span class=\"EOP SCXW156847865 BCX0\" data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335557856&quot;:16777215,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335572071&quot;:4,&quot;335572072&quot;:4,&quot;335572073&quot;:0,&quot;335572075&quot;:4,&quot;335572076&quot;:4,&quot;335572077&quot;:0,&quot;335572079&quot;:4,&quot;335572080&quot;:4,&quot;335572081&quot;:0,&quot;335572083&quot;:4,&quot;335572084&quot;:4,&quot;335572085&quot;:0,&quot;469789798&quot;:&quot;nil&quot;,&quot;469789802&quot;:&quot;nil&quot;,&quot;469789806&quot;:&quot;nil&quot;,&quot;469789810&quot;:&quot;nil&quot;}\">\u00a0<\/span><\/h3>\n\t<div id=\"gap-1352331874\" class=\"gap-element clearfix\" style=\"display:block; height:auto;\">\n\t\t\n<style>\n#gap-1352331874 {\n  padding-top: 10px;\n}\n<\/style>\n\t<\/div>\n\t\n<ul>\n<li><b><span data-contrast=\"none\">KI\/ML<\/span><\/b><span data-contrast=\"none\">: K\u00fcnstliche Intelligenz \/ Maschinelles Lernen \u2013 Systeme, die Muster aus Daten lernen, anstatt einer festen Programmierung zu folgen.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335557856&quot;:16777215,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335572071&quot;:4,&quot;335572072&quot;:4,&quot;335572073&quot;:0,&quot;335572075&quot;:4,&quot;335572076&quot;:4,&quot;335572077&quot;:0,&quot;335572079&quot;:4,&quot;335572080&quot;:4,&quot;335572081&quot;:0,&quot;335572083&quot;:4,&quot;335572084&quot;:4,&quot;335572085&quot;:0,&quot;469789798&quot;:&quot;nil&quot;,&quot;469789802&quot;:&quot;nil&quot;,&quot;469789806&quot;:&quot;nil&quot;,&quot;469789810&quot;:&quot;nil&quot;}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"none\">Statisches Modell<\/span><\/b><span data-contrast=\"none\">: Ein KI-Modell, das sich nach dem Training nicht \u00e4ndert; erforderlich in kritischen GMP-Anwendungen.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335557856&quot;:16777215,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335572071&quot;:4,&quot;335572072&quot;:4,&quot;335572073&quot;:0,&quot;335572075&quot;:4,&quot;335572076&quot;:4,&quot;335572077&quot;:0,&quot;335572079&quot;:4,&quot;335572080&quot;:4,&quot;335572081&quot;:0,&quot;335572083&quot;:4,&quot;335572084&quot;:4,&quot;335572085&quot;:0,&quot;469789798&quot;:&quot;nil&quot;,&quot;469789802&quot;:&quot;nil&quot;,&quot;469789806&quot;:&quot;nil&quot;,&quot;469789810&quot;:&quot;nil&quot;}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"none\">Dynamisches Modell<\/span><\/b><span data-contrast=\"none\">: Kontinuierlich lernende\/adaptive KI-Systeme; nicht zul\u00e4ssig im kritischen GMP-Einsatz.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335557856&quot;:16777215,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335572071&quot;:4,&quot;335572072&quot;:4,&quot;335572073&quot;:0,&quot;335572075&quot;:4,&quot;335572076&quot;:4,&quot;335572077&quot;:0,&quot;335572079&quot;:4,&quot;335572080&quot;:4,&quot;335572081&quot;:0,&quot;335572083&quot;:4,&quot;335572084&quot;:4,&quot;335572085&quot;:0,&quot;469789798&quot;:&quot;nil&quot;,&quot;469789802&quot;:&quot;nil&quot;,&quot;469789806&quot;:&quot;nil&quot;,&quot;469789810&quot;:&quot;nil&quot;}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"none\">Deterministische Ausgabe<\/span><\/b><span data-contrast=\"none\">: Vorhersehbare, wiederholbare Ergebnisse f\u00fcr eine gegebene Eingabe.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335557856&quot;:16777215,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335572071&quot;:4,&quot;335572072&quot;:4,&quot;335572073&quot;:0,&quot;335572075&quot;:4,&quot;335572076&quot;:4,&quot;335572077&quot;:0,&quot;335572079&quot;:4,&quot;335572080&quot;:4,&quot;335572081&quot;:0,&quot;335572083&quot;:4,&quot;335572084&quot;:4,&quot;335572085&quot;:0,&quot;469789798&quot;:&quot;nil&quot;,&quot;469789802&quot;:&quot;nil&quot;,&quot;469789806&quot;:&quot;nil&quot;,&quot;469789810&quot;:&quot;nil&quot;}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"none\">Probabilistische Ausgabe<\/span><\/b><span data-contrast=\"none\">: Variable Ergebnisse auch bei identischer Eingabe; in kritischen Anwendungsf\u00e4llen nicht zul\u00e4ssig.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335557856&quot;:16777215,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335572071&quot;:4,&quot;335572072&quot;:4,&quot;335572073&quot;:0,&quot;335572075&quot;:4,&quot;335572076&quot;:4,&quot;335572077&quot;:0,&quot;335572079&quot;:4,&quot;335572080&quot;:4,&quot;335572081&quot;:0,&quot;335572083&quot;:4,&quot;335572084&quot;:4,&quot;335572085&quot;:0,&quot;469789798&quot;:&quot;nil&quot;,&quot;469789802&quot;:&quot;nil&quot;,&quot;469789806&quot;:&quot;nil&quot;,&quot;469789810&quot;:&quot;nil&quot;}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"none\">Generative KI \/ LLM<\/span><\/b><span data-contrast=\"none\">: Inhaltsgenerierende Modelle (z. B. Text- oder Bildgeneratoren), die probabilistisch arbeiten.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335557856&quot;:16777215,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335572071&quot;:4,&quot;335572072&quot;:4,&quot;335572073&quot;:0,&quot;335572075&quot;:4,&quot;335572076&quot;:4,&quot;335572077&quot;:0,&quot;335572079&quot;:4,&quot;335572080&quot;:4,&quot;335572081&quot;:0,&quot;335572083&quot;:4,&quot;335572084&quot;:4,&quot;335572085&quot;:0,&quot;469789798&quot;:&quot;nil&quot;,&quot;469789802&quot;:&quot;nil&quot;,&quot;469789806&quot;:&quot;nil&quot;,&quot;469789810&quot;:&quot;nil&quot;}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"none\">HITL (Human-In-The-Loop)<\/span><\/b><span data-contrast=\"none\">: \u00dcberwachungsmechanismus, der eine menschliche \u00dcberpr\u00fcfung oder Validierung von KI-Entscheidungen bei nicht kritischen Anwendungen erfordert.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335557856&quot;:16777215,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335572071&quot;:4,&quot;335572072&quot;:4,&quot;335572073&quot;:0,&quot;335572075&quot;:4,&quot;335572076&quot;:4,&quot;335572077&quot;:0,&quot;335572079&quot;:4,&quot;335572080&quot;:4,&quot;335572081&quot;:0,&quot;335572083&quot;:4,&quot;335572084&quot;:4,&quot;335572085&quot;:0,&quot;469789798&quot;:&quot;nil&quot;,&quot;469789802&quot;:&quot;nil&quot;,&quot;469789806&quot;:&quot;nil&quot;,&quot;469789810&quot;:&quot;nil&quot;}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"none\">Erkl\u00e4rbarkeit<\/span><\/b><span data-contrast=\"none\">: F\u00e4higkeit, KI-Entscheidungen und -Logik zu verstehen und zu interpretieren.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335557856&quot;:16777215,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335572071&quot;:4,&quot;335572072&quot;:4,&quot;335572073&quot;:0,&quot;335572075&quot;:4,&quot;335572076&quot;:4,&quot;335572077&quot;:0,&quot;335572079&quot;:4,&quot;335572080&quot;:4,&quot;335572081&quot;:0,&quot;335572083&quot;:4,&quot;335572084&quot;:4,&quot;335572085&quot;:0,&quot;469789798&quot;:&quot;nil&quot;,&quot;469789802&quot;:&quot;nil&quot;,&quot;469789806&quot;:&quot;nil&quot;,&quot;469789810&quot;:&quot;nil&quot;}\">\u00a0<\/span><\/li>\n<li><b><span data-contrast=\"none\">Testdatenunabh\u00e4ngigkeit<\/span><\/b><span data-contrast=\"none\">: Verwendung von Datens\u00e4tzen getrennt von Trainingsdaten, um die KI-Leistung objektiv zu validieren.<\/span><span data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335557856&quot;:16777215,&quot;335559738&quot;:210,&quot;335559739&quot;:210,&quot;335572071&quot;:4,&quot;335572072&quot;:4,&quot;335572073&quot;:0,&quot;335572075&quot;:4,&quot;335572076&quot;:4,&quot;335572077&quot;:0,&quot;335572079&quot;:4,&quot;335572080&quot;:4,&quot;335572081&quot;:0,&quot;335572083&quot;:4,&quot;335572084&quot;:4,&quot;335572085&quot;:0,&quot;469789798&quot;:&quot;nil&quot;,&quot;469789802&quot;:&quot;nil&quot;,&quot;469789806&quot;:&quot;nil&quot;,&quot;469789810&quot;:&quot;nil&quot;}\">\u00a0<\/span><\/li>\n<\/ul>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\n\t\n<\/div>\n<div class=\"row\"  id=\"row-1889510781\">\n\n\t<div id=\"col-2121267131\" class=\"col small-12 large-12\"  >\n\t\t\t\t<div class=\"col-inner\"  >\n\t\t\t\n\t\t\t\n<div class=\"is-divider divider clearfix\" style=\"margin-top:2em;margin-bottom:2em;max-width:20%;height:1px;background-color:#a5acd9;\"><\/div>\n<h3>FAQs<\/h3>\n\t\t<\/div>\n\t\t\t\t\n<style>\n#col-2121267131 > .col-inner {\n  margin: 0px 0px -10px 0px;\n}\n<\/style>\n\t<\/div>\n\n\t\n\n\t<div id=\"col-1245383229\" class=\"col small-12 large-12\"  >\n\t\t\t\t<div class=\"col-inner\"  >\n\t\t\t\n\t\t\t\n\t<div id=\"text-1418627880\" class=\"text\">\n\t\t\n<h4><span class=\"TextRun MacChromeBold SCXW43675964 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW43675964 BCX0\">Was ist Anhang 22 im PIC\/S-GMP-Leitfaden?<\/span><\/span><\/h4>\n\t\t\n<style>\n#text-1418627880 {\n  line-height: 2.1;\n}\n<\/style>\n\t<\/div>\n\t\n<p><span class=\"TextRun SCXW65768994 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW65768994 BCX0\">Anhang 22 ist eine neue Richtlinie von PIC\/S, die die regulatorischen Erwartungen f\u00fcr den Einsatz von KI und ML in der GMP-regulierten Arzneimittelherstellung festlegt.<\/span><\/span><\/p>\n\t\t<\/div>\n\t\t\t\t\n<style>\n#col-1245383229 > .col-inner {\n  margin: 0px 0px -20px 0px;\n}\n<\/style>\n\t<\/div>\n\n\t\n\n\t<div id=\"col-1657813724\" class=\"col small-12 large-12\"  >\n\t\t\t\t<div class=\"col-inner\"  >\n\t\t\t\n\t\t\t\n\t<div id=\"text-3869216007\" class=\"text\">\n\t\t\n<h4><span class=\"TextRun MacChromeBold SCXW93685708 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW93685708 BCX0\">Sind dynamische KI- oder generative KI-Modelle in GMP-Anwendungen zul\u00e4ssig?<\/span><\/span><span class=\"LineBreakBlob BlobObject DragDrop SCXW93685708 BCX0\"><span class=\"SCXW93685708 BCX0\">\u00a0<\/span><br class=\"SCXW93685708 BCX0\" \/><\/span><\/h4>\n\t\t\n<style>\n#text-3869216007 {\n  line-height: 2.1;\n}\n<\/style>\n\t<\/div>\n\t\n<p><span class=\"TextRun SCXW9051628 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW9051628 BCX0\">Dynamische, adaptive oder probabilistische KI-Modelle sind in kritischen GMP-Anwendungen verboten. Generative KI und LLMs d\u00fcrfen nur in nicht-kritischen Anwendungen mit Human-in-the-Loop-\u00dcberwachung eingesetzt werden.<\/span><\/span><span class=\"EOP SCXW9051628 BCX0\" data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n\t\t<\/div>\n\t\t\t\t\n<style>\n#col-1657813724 > .col-inner {\n  margin: 0px 0px -20px 0px;\n}\n<\/style>\n\t<\/div>\n\n\t\n\n\t<div id=\"col-2105146584\" class=\"col small-12 large-12\"  >\n\t\t\t\t<div class=\"col-inner\"  >\n\t\t\t\n\t\t\t\n\t<div id=\"text-858741880\" class=\"text\">\n\t\t\n<h4><span class=\"LineBreakBlob BlobObject DragDrop SCXW93685708 BCX0\"><span class=\"TextRun MacChromeBold SCXW71618274 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW71618274 BCX0\">Warum konzentriert sich Anhang 22 auf statische und deterministische KI-Modelle?<\/span><\/span><span class=\"LineBreakBlob BlobObject DragDrop SCXW71618274 BCX0\"><span class=\"SCXW71618274 BCX0\">\u00a0<\/span><br class=\"SCXW71618274 BCX0\" \/><\/span><\/span><\/h4>\n\t\t\n<style>\n#text-858741880 {\n  line-height: 2.1;\n}\n<\/style>\n\t<\/div>\n\t\n<p><span class=\"TextRun SCXW75775714 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW75775714 BCX0\">Statische und deterministische Modelle liefern vorhersehbare, wiederholbare Ergebnisse und reduzieren Risiken f\u00fcr die Patientensicherheit, Produktqualit\u00e4t und Datenintegrit\u00e4t.<\/span><\/span><span class=\"EOP SCXW75775714 BCX0\" data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;335559738&quot;:240,&quot;335559739&quot;:240}\">\u00a0<\/span><\/p>\n\t\t<\/div>\n\t\t\t\t\n<style>\n#col-2105146584 > .col-inner {\n  margin: 0px 0px -20px 0px;\n}\n<\/style>\n\t<\/div>\n\n\t\n\n\t<div id=\"col-539271670\" class=\"col small-12 large-12\"  >\n\t\t\t\t<div class=\"col-inner\"  >\n\t\t\t\n\t\t\t\n\t<div id=\"text-3900849588\" class=\"text\">\n\t\t\n<h4><span class=\"LineBreakBlob BlobObject DragDrop SCXW93685708 BCX0\"><span class=\"LineBreakBlob BlobObject DragDrop SCXW71618274 BCX0\"><span class=\"NormalTextRun SCXW110541706 BCX0\">Wie kann\u00a0<\/span><span class=\"NormalTextRun SpellingErrorV2Themed SCXW110541706 BCX0\">RegASK<\/span><span class=\"NormalTextRun SCXW110541706 BCX0\">\u00a0Unternehmen helfen\u00a0<\/span><span class=\"NormalTextRun SCXW110541706 BCX0\">entsprechen<\/span><span class=\"NormalTextRun SCXW110541706 BCX0\">\u00a0Anhang 22?<\/span><\/span><\/span><\/h4>\n\t\t\n<style>\n#text-3900849588 {\n  line-height: 2.1;\n}\n<\/style>\n\t<\/div>\n\t\n<p><a class=\"Hyperlink SCXW12507906 BCX0\" href=\"https:\/\/regask.com\/de\/\" target=\"_blank\" rel=\"noreferrer noopener\"><span class=\"TextRun Underlined SCXW12507906 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW12507906 BCX0\" data-ccp-charstyle=\"Hyperlink\">RegASK<\/span><\/span><\/a><span class=\"TextRun SCXW12507906 BCX0\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW12507906 BCX0\">\u00a0bietet KI-gesteuerte regulatorische Intelligenz und Workflow-Orchestrierung f\u00fcr\u00a0<\/span><span class=\"NormalTextRun SCXW12507906 BCX0\">Monitor<\/span><span class=\"NormalTextRun SCXW12507906 BCX0\">\u00a0Neue Anforderungen wie Anhang 22, gleichen Sie diese mit bestehenden Unternehmensprozessen ab und zeigen Sie Compliance-L\u00fccken auf. Unsere Plattform unterst\u00fctzt strukturierte Dokumentation, Validierungsworkflows und Expertenwissen und hilft Teams, sich schneller anzupassen und regulatorische Risiken zu reduzieren.<\/span><\/span><\/p>\n\t\t<\/div>\n\t\t\t\t\n<style>\n#col-539271670 > .col-inner {\n  margin: 0px 0px -20px 0px;\n}\n<\/style>\n\t<\/div>\n\n\t\n<\/div>","protected":false},"excerpt":{"rendered":"<p>Am 7. Juli 2025 ver\u00f6ffentlichte das Pharmaceutical Inspection Co-operation Scheme (PIC\/S) Anhang 22 zum PIC\/S GMP-Leitfaden. Dieser neue Anhang enth\u00e4lt detaillierte regulatorische Anforderungen f\u00fcr den Einsatz von K\u00fcnstlicher Intelligenz (KI) und Maschinellem Lernen (ML) in GMP-regulierten Umgebungen. Die Aktualisierung schlie\u00dft L\u00fccken in Anhang 11, der die sich entwickelnden KI\/ML-Technologien nicht vollst\u00e4ndig abdeckt. Ziel ist es, eine sichere, zuverl\u00e4ssige und transparente KI-Integration zu gew\u00e4hrleisten\u2026<\/p>","protected":false},"author":32,"featured_media":8663,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[33],"tags":[],"class_list":["post-8662","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v27.6 (Yoast SEO v27.6) - https:\/\/yoast.com\/product\/yoast-seo-premium-wordpress\/ -->\n<title>Decoding the New PIC\/S Annex 22 for Regulatory Teams | RegASK<\/title>\n<meta name=\"description\" content=\"Find out What the New PIC\/S Annex 22 Means for Compliance Teams regarding compliance with AI and ML standards in pharmaceuticals.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/regask.com\/de\/ki-gmp-dekodierung-neuer-bilder-anhang-22-regulierungsteams\/\" \/>\n<meta property=\"og:locale\" content=\"de_DE\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"AI in GMP: Decoding the New PIC\/S Annex 22 for Regulatory Teams\" \/>\n<meta property=\"og:description\" content=\"Find out What the New PIC\/S Annex 22 Means for Compliance Teams regarding compliance with AI and ML standards in pharmaceuticals.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/regask.com\/de\/ki-gmp-dekodierung-neuer-bilder-anhang-22-regulierungsteams\/\" \/>\n<meta property=\"og:site_name\" content=\"RegASK\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/RegASKRegulatoryAffairs\/\" \/>\n<meta property=\"article:published_time\" content=\"2025-09-11T03:55:50+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2025-10-15T06:47:31+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/regask.com\/wp-content\/uploads\/2025\/09\/ai-gmp-new-pics-annex-22-means-compliance-teams.png\" \/>\n\t<meta property=\"og:image:width\" content=\"800\" \/>\n\t<meta property=\"og:image:height\" content=\"600\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Hannah Nordin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@ASKRegASK\" \/>\n<meta name=\"twitter:site\" content=\"@ASKRegASK\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Hannah Nordin\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"4\u00a0Minuten\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/regask.com\\\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/regask.com\\\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\\\/\"},\"author\":{\"name\":\"Hannah Nordin\",\"@id\":\"https:\\\/\\\/regask.com\\\/#\\\/schema\\\/person\\\/3a8aaee2c7585284f97ec9e038e36b65\"},\"headline\":\"AI in GMP: Decoding the New PIC\\\/S Annex 22 for Regulato&hellip;\",\"datePublished\":\"2025-09-11T03:55:50+00:00\",\"dateModified\":\"2025-10-15T06:47:31+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/regask.com\\\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\\\/\"},\"wordCount\":867,\"publisher\":{\"@id\":\"https:\\\/\\\/regask.com\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/regask.com\\\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/regask.com\\\/wp-content\\\/uploads\\\/2025\\\/09\\\/ai-gmp-new-pics-annex-22-means-compliance-teams.png\",\"articleSection\":[\"Regulatory News\"],\"inLanguage\":\"de\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/regask.com\\\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\\\/\",\"url\":\"https:\\\/\\\/regask.com\\\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\\\/\",\"name\":\"Decoding the New PIC\\\/S Annex 22 for Regulatory Teams | RegASK\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/regask.com\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/regask.com\\\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/regask.com\\\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/regask.com\\\/wp-content\\\/uploads\\\/2025\\\/09\\\/ai-gmp-new-pics-annex-22-means-compliance-teams.png\",\"datePublished\":\"2025-09-11T03:55:50+00:00\",\"dateModified\":\"2025-10-15T06:47:31+00:00\",\"description\":\"Find out What the New PIC\\\/S Annex 22 Means for Compliance Teams regarding compliance with AI and ML standards in pharmaceuticals.\",\"inLanguage\":\"de\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/regask.com\\\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"de\",\"@id\":\"https:\\\/\\\/regask.com\\\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\\\/#primaryimage\",\"url\":\"https:\\\/\\\/regask.com\\\/wp-content\\\/uploads\\\/2025\\\/09\\\/ai-gmp-new-pics-annex-22-means-compliance-teams.png\",\"contentUrl\":\"https:\\\/\\\/regask.com\\\/wp-content\\\/uploads\\\/2025\\\/09\\\/ai-gmp-new-pics-annex-22-means-compliance-teams.png\",\"width\":800,\"height\":600,\"caption\":\"Ai Gmp New Pics Annex 22 Means Compliance Teams\"},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/regask.com\\\/#website\",\"url\":\"https:\\\/\\\/regask.com\\\/\",\"name\":\"RegASK\",\"description\":\"The Regulatory AI Operating System for High-Performing Teams\",\"publisher\":{\"@id\":\"https:\\\/\\\/regask.com\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/regask.com\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"de\"},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/regask.com\\\/#\\\/schema\\\/person\\\/3a8aaee2c7585284f97ec9e038e36b65\",\"name\":\"Hannah Nordin\",\"sameAs\":[\"https:\\\/\\\/regask.com\"]}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Entschl\u00fcsselung des neuen PIC\/S-Anhangs 22 f\u00fcr Regulierungsteams | RegASK","description":"Informieren Sie sich, was der neue PIC\/S-Anhang 22 f\u00fcr Compliance-Teams hinsichtlich der Einhaltung von KI- und ML-Standards in der Pharmaindustrie bedeutet.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/regask.com\/de\/ki-gmp-dekodierung-neuer-bilder-anhang-22-regulierungsteams\/","og_locale":"de_DE","og_type":"article","og_title":"AI in GMP: Decoding the New PIC\/S Annex 22 for Regulatory Teams","og_description":"Find out What the New PIC\/S Annex 22 Means for Compliance Teams regarding compliance with AI and ML standards in pharmaceuticals.","og_url":"https:\/\/regask.com\/de\/ki-gmp-dekodierung-neuer-bilder-anhang-22-regulierungsteams\/","og_site_name":"RegASK","article_publisher":"https:\/\/www.facebook.com\/RegASKRegulatoryAffairs\/","article_published_time":"2025-09-11T03:55:50+00:00","article_modified_time":"2025-10-15T06:47:31+00:00","og_image":[{"width":800,"height":600,"url":"https:\/\/regask.com\/wp-content\/uploads\/2025\/09\/ai-gmp-new-pics-annex-22-means-compliance-teams.png","type":"image\/png"}],"author":"Hannah Nordin","twitter_card":"summary_large_image","twitter_creator":"@ASKRegASK","twitter_site":"@ASKRegASK","twitter_misc":{"Written by":"Hannah Nordin","Est. reading time":"4\u00a0Minuten"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/regask.com\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/#article","isPartOf":{"@id":"https:\/\/regask.com\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/"},"author":{"name":"Hannah Nordin","@id":"https:\/\/regask.com\/#\/schema\/person\/3a8aaee2c7585284f97ec9e038e36b65"},"headline":"AI in GMP: Decoding the New PIC\/S Annex 22 for Regulato&hellip;","datePublished":"2025-09-11T03:55:50+00:00","dateModified":"2025-10-15T06:47:31+00:00","mainEntityOfPage":{"@id":"https:\/\/regask.com\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/"},"wordCount":867,"publisher":{"@id":"https:\/\/regask.com\/#organization"},"image":{"@id":"https:\/\/regask.com\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/#primaryimage"},"thumbnailUrl":"https:\/\/regask.com\/wp-content\/uploads\/2025\/09\/ai-gmp-new-pics-annex-22-means-compliance-teams.png","articleSection":["Regulatory News"],"inLanguage":"de"},{"@type":"WebPage","@id":"https:\/\/regask.com\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/","url":"https:\/\/regask.com\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/","name":"Entschl\u00fcsselung des neuen PIC\/S-Anhangs 22 f\u00fcr Regulierungsteams | RegASK","isPartOf":{"@id":"https:\/\/regask.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/regask.com\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/#primaryimage"},"image":{"@id":"https:\/\/regask.com\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/#primaryimage"},"thumbnailUrl":"https:\/\/regask.com\/wp-content\/uploads\/2025\/09\/ai-gmp-new-pics-annex-22-means-compliance-teams.png","datePublished":"2025-09-11T03:55:50+00:00","dateModified":"2025-10-15T06:47:31+00:00","description":"Informieren Sie sich, was der neue PIC\/S-Anhang 22 f\u00fcr Compliance-Teams hinsichtlich der Einhaltung von KI- und ML-Standards in der Pharmaindustrie bedeutet.","inLanguage":"de","potentialAction":[{"@type":"ReadAction","target":["https:\/\/regask.com\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/"]}]},{"@type":"ImageObject","inLanguage":"de","@id":"https:\/\/regask.com\/ai-gmp-decoding-new-pics-annex-22-regulatory-teams\/#primaryimage","url":"https:\/\/regask.com\/wp-content\/uploads\/2025\/09\/ai-gmp-new-pics-annex-22-means-compliance-teams.png","contentUrl":"https:\/\/regask.com\/wp-content\/uploads\/2025\/09\/ai-gmp-new-pics-annex-22-means-compliance-teams.png","width":800,"height":600,"caption":"Ai Gmp New Pics Annex 22 Means Compliance Teams"},{"@type":"WebSite","@id":"https:\/\/regask.com\/#website","url":"https:\/\/regask.com\/","name":"RegASK","description":"Das regulatorische KI-Betriebssystem f\u00fcr leistungsstarke Teams","publisher":{"@id":"https:\/\/regask.com\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/regask.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"de"},{"@type":"Person","@id":"https:\/\/regask.com\/#\/schema\/person\/3a8aaee2c7585284f97ec9e038e36b65","name":"Hannah Nordin","sameAs":["https:\/\/regask.com"]}]}},"_links":{"self":[{"href":"https:\/\/regask.com\/de\/wp-json\/wp\/v2\/posts\/8662","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/regask.com\/de\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/regask.com\/de\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/regask.com\/de\/wp-json\/wp\/v2\/users\/32"}],"replies":[{"embeddable":true,"href":"https:\/\/regask.com\/de\/wp-json\/wp\/v2\/comments?post=8662"}],"version-history":[{"count":0,"href":"https:\/\/regask.com\/de\/wp-json\/wp\/v2\/posts\/8662\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/regask.com\/de\/wp-json\/wp\/v2\/media\/8663"}],"wp:attachment":[{"href":"https:\/\/regask.com\/de\/wp-json\/wp\/v2\/media?parent=8662"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/regask.com\/de\/wp-json\/wp\/v2\/categories?post=8662"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/regask.com\/de\/wp-json\/wp\/v2\/tags?post=8662"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}