<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Phillip Inzaghi on Medium]]></title>
        <description><![CDATA[Stories by Phillip Inzaghi on Medium]]></description>
        <link>https://medium.com/@inzaghi.phil?source=rss-38c1d6aefa3f------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*JsZ-fhjFTTjkB1rE</url>
            <title>Stories by Phillip Inzaghi on Medium</title>
            <link>https://medium.com/@inzaghi.phil?source=rss-38c1d6aefa3f------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 06:36:11 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@inzaghi.phil/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[เขาได้ยินเรื่องราวของเพื่อนสนิทที่เป็นลูกพี่ลูกน้องของภรรยามานานหลายปีก่อนจะได้พบหน้ากันจริง ๆ…]]></title>
            <link>https://medium.com/@inzaghi.phil/%E0%B9%80%E0%B8%82%E0%B8%B2%E0%B9%84%E0%B8%94%E0%B9%89%E0%B8%A2%E0%B8%B4%E0%B8%99%E0%B9%80%E0%B8%A3%E0%B8%B7%E0%B9%88%E0%B8%AD%E0%B8%87%E0%B8%A3%E0%B8%B2%E0%B8%A7%E0%B8%82%E0%B8%AD%E0%B8%87%E0%B9%80%E0%B8%9E%E0%B8%B7%E0%B9%88%E0%B8%AD%E0%B8%99%E0%B8%AA%E0%B8%99%E0%B8%B4%E0%B8%97%E0%B8%97%E0%B8%B5%E0%B9%88%E0%B9%80%E0%B8%9B%E0%B9%87%E0%B8%99%E0%B8%A5%E0%B8%B9%E0%B8%81%E0%B8%9E%E0%B8%B5%E0%B9%88%E0%B8%A5%E0%B8%B9%E0%B8%81%E0%B8%99%E0%B9%89%E0%B8%AD%E0%B8%87%E0%B8%82%E0%B8%AD%E0%B8%87%E0%B8%A0%E0%B8%A3%E0%B8%A3%E0%B8%A2%E0%B8%B2%E0%B8%A1%E0%B8%B2%E0%B8%99%E0%B8%B2%E0%B8%99%E0%B8%AB%E0%B8%A5%E0%B8%B2%E0%B8%A2%E0%B8%9B%E0%B8%B5%E0%B8%81%E0%B9%88%E0%B8%AD%E0%B8%99%E0%B8%88%E0%B8%B0%E0%B9%84%E0%B8%94%E0%B9%89%E0%B8%9E%E0%B8%9A%E0%B8%AB%E0%B8%99%E0%B9%89%E0%B8%B2%E0%B8%81%E0%B8%B1%E0%B8%99%E0%B8%88%E0%B8%A3%E0%B8%B4%E0%B8%87-%E0%B9%86-23a8071b4119?source=rss-38c1d6aefa3f------2</link>
            <guid isPermaLink="false">https://medium.com/p/23a8071b4119</guid>
            <category><![CDATA[love]]></category>
            <category><![CDATA[forbiddenlove]]></category>
            <category><![CDATA[sexuality]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[cheating]]></category>
            <dc:creator><![CDATA[Phillip Inzaghi]]></dc:creator>
            <pubDate>Wed, 09 Sep 2026 18:53:01 GMT</pubDate>
            <atom:updated>2026-09-11T09:51:40.369Z</atom:updated>
            <content:encoded><![CDATA[<p>เขาได้ยินเรื่องราวของเพื่อนสนิทที่เป็นลูกพี่ลูกน้องของภรรยามานานหลายปีก่อนจะได้พบหน้ากันจริง ๆ มีบางวูบที่เขาแอบคิดเล่น ๆ ว่า ที่ผ่านมาคลาดกันตลอดนั้น เป็นเพราะภรรยาจงใจไม่ยอมให้ทั้งคู่ได้เจอกันหรือเปล่า<br>​ทว่าเมื่อได้พบกันเข้าจริง มันกลับไม่มีแม้แต่เศษเสี้ยวของการหยั่งเชิงทอดสะพาน ไม่มีใครคิดจะสร้างภาพให้อีกฝ่ายประทับใจ ด้วยสถานะที่ขีดคั่นไว้อย่างรัดกุมตั้งแต่แรกเริ่ม เขาคือชายหนุ่มที่กำลังจะก้าวเข้ามาเป็นคนรักของคนที่เธอหวงแหน ส่วนเธอคือเพื่อนสนิทที่คอยจับตาดูว่าผู้ชายคนนี้คู่ควรกับลูกพี่ลูกน้องของเธอจริงไหม<br>​เส้นแบ่งระหว่างเขากับเธอจึงถูกลากไว้อย่างชัดเจน ก่อนที่มิตรภาพจะได้เริ่มต้นเสียด้วยซ้ำ<br>​และบางที อาจเป็นเพราะเส้นแบ่งนั้นเอง ที่ทำให้ต่างฝ่ายต่างได้เห็นเนื้อแท้ของกันและกันโดยไม่มีอะไรบดบัง<br>​พวกเขาเหมือนกันอย่างน่าประหลาด<br>​เป็นพวกนักวางแผน เป็นคนคอยจัดการและแบกรับปัญหาของคนอื่นไว้บนบ่าอยู่เสมอ ทั้งคู่คุ้นชินกับการมองทะลุว่าคนรอบข้างต้องการอะไรแล้วลงมือทำให้ลุล่วง และต่างก็มีทิฐิเกินกว่าจะยอมปริปากว่า ลึก ๆ แล้วตัวเองก็โหยหาใครสักคนมาคอยดูแลเหมือนกัน<br>​ตามหลักแล้ว คนหัวแข็งและมีเหลี่ยมมุมคล้ายกันขนาดนี้ อยู่ใกล้กันคงมีแต่จะปะทะ ทว่าพวกเขากลับเข้าขากันได้อย่างไม่น่าเชื่อ<br>​ความเข้ากันได้นี้เด่นชัดที่สุดในคราวที่ต้องเตรียมงานแต่งงาน<br>​เขาเป็นนายทุนและผู้ควบคุมงาน ส่วนเธอกลายเป็นทั้งเพื่อนเจ้าสาว ผู้ประสานงาน และฝ่ายดับเพลิงเฉพาะกิจไปโดยปริยาย ตลอดหลายสัปดาห์ก่อนวันงาน ข้อความระหว่างทั้งคู่แทบไม่เคยขาดสาย มีการตัดสินใจนับร้อยและสารพัดปัญหาจุกจิกให้คลี่คลาย บ่อยครั้งที่คนหนึ่งเพิ่งเริ่มเอ่ยปาก อีกฝ่ายก็รู้ใจทันทีว่าต้องทำอย่างไรต่อ<br>​งานแต่งงานออกมาสมบูรณ์แบบไร้ที่ติทว่าคำตอบที่ผุดขึ้นมากลับทำให้เขาประหลาดใจ</p><p>สิ่งที่เขาจินตนาการถึง ไม่ได้เริ่มต้นจากเรื่องบนเตียง</p><p>เขาอยากถ่ายรูปเธอ</p><p>เธอเป็นคนที่ดูดีอย่างเป็นธรรมชาติเมื่ออยู่หน้ากล้อง ส่วนเขาก็รู้ว่าต้องทำอย่างไรเมื่ออยู่หลังเลนส์ เขาเคยนึกเล่น ๆ อยู่หลายครั้งว่า หากวันหนึ่งได้มีโอกาสถ่ายภาพเธอจริง ๆ ภาพเหล่านั้นจะออกมาเป็นอย่างไร</p><p>ไม่ใช่ภาพธรรมดาที่ใคร ๆ ก็ถ่ายได้</p><p>แต่เป็นภาพที่ต้องอาศัยความไว้ใจ</p><p>ภาพที่ค่อย ๆ ลดระยะห่างระหว่างคนหน้ากล้องกับคนหลังกล้องลงทีละน้อย จากความสวยงามที่เปิดเผยต่อสายตาคนทั่วไป ไปสู่ความงามในมุมที่เปราะบาง เป็นส่วนตัว และไม่ใช่สิ่งที่เธอจะยอมให้ใครเห็นได้ง่าย ๆ</p><p>บางครั้งเขาก็ปล่อยให้จินตนาการเดินไปไกลกว่านั้น</p><p>ไม่ใช่เพราะเขาควบคุมตัวเองไม่ได้ แต่เพราะเขาอยากรู้ว่า ถ้าคนสองคนที่เข้าใจกันมากขนาดนี้ ได้อยู่ด้วยกันโดยไม่มีใครอื่น ไม่มีบทบาท ไม่มีหน้าที่ และไม่มีสายตาของคนรอบข้างคอยกำหนดว่าพวกเขาควรเป็นใคร มันจะรู้สึกอย่างไร</p><p>ตอนที่เขาเคยพูดทีเล่นทีจริงว่า</p><p>“สามชั่วโมงคงไม่พอ”</p><p>ความจริงคือ เขาไม่ได้ล้อเล่นทั้งหมด</p><p>สามชั่วโมงคงไม่พอสำหรับสิ่งที่เขาจินตนาการไว้</p><p>เพราะในความคิดของเขา มันไม่เคยเป็นเรื่องของความรีบร้อน</p><p>สิ่งที่ดึงดูดเขาคือความคิดที่จะมอบเวลาและความสนใจทั้งหมดให้ผู้หญิงคนหนึ่งที่ใช้ชีวิตส่วนใหญ่คอยมอบสิ่งเหล่านั้นให้คนอื่น</p><p>ผู้หญิงที่คอยวางแผน</p><p>คอยจัดการ</p><p>คอยดูแล</p><p>คอยสังเกตว่าทุกคนต้องการอะไร</p><p>แต่แทบไม่เคยยอมพูดออกมาตรง ๆ ว่าตัวเองต้องการอะไร</p><p>ในจินตนาการของเขา เพียงช่วงเวลาหนึ่ง เธอไม่จำเป็นต้องทำอะไรเลย</p><p>ไม่ต้องรับผิดชอบใคร</p><p>ไม่ต้องแก้ปัญหา</p><p>ไม่ต้องคิดล่วงหน้า</p><p>ไม่ต้องเป็นคนเข้มแข็งที่ทุกคนคุ้นเคย</p><p>เธอเพียงแค่เป็นตัวเอง และปล่อยให้ใครสักคนเป็นฝ่ายใส่ใจเธอบ้าง</p><p>บางครั้งจินตนาการนั้นก็พาเขาไปถึงคำถามที่อันตรายกว่าเดิม</p><p>ถ้าความใกล้ชิดนั้นพาพวกเขาข้ามเส้นไปจริง ๆ ล่ะ?</p><p>เขารู้คำตอบของตัวเองดีพอที่จะไม่จำเป็นต้องโกหกตัวเอง</p><p>เขาอยากรู้</p><p>อยากรู้ว่าการได้ใกล้ชิดกับเธอจริง ๆ จะเป็นอย่างไร อยากรู้ว่าความเข้าใจกันที่เกิดขึ้นมาตลอดหลายปีจะยังคงอยู่หรือไม่ เมื่อไม่มีมุกตลก ไม่มีข้อความสองแง่สองง่าม และไม่มีระยะปลอดภัยของหน้าจอคั่นกลาง</p><p>แต่แม้ในจินตนาการ เขาก็ไม่เคยเห็นชีวิตอีกชีวิตหนึ่งรออยู่หลังจากนั้น</p><p>เขายังคงรักภรรยา</p><p>ไม่เคยคิดจะทิ้งเธอ</p><p>ไม่เคยจินตนาการถึงบ้านอีกหลัง ชีวิตคู่แบบใหม่ หรือการเอาผู้หญิงคนหนึ่งมาแทนที่อีกคน</p><p>สิ่งที่เขาจินตนาการจึงมีขอบเขตของมันเอง</p><p>ช่วงเวลาหนึ่งที่เป็นของคนสองคน</p><p>ประสบการณ์หนึ่ง</p><p>ความลับหนึ่ง</p><p>แล้วทุกอย่างก็กลับไปอยู่ในที่ของมัน</p><p>ในจินตนาการของเขา หลังจากนั้นทั้งคู่ยังสามารถกลับบ้านไปหาคนรักของตัวเองได้</p><p>บางทีในเย็นวันเดียวกัน คนทั้งสี่อาจยังออกไปกินข้าวด้วยกันเหมือนทุกครั้ง</p><p>ไม่มีใครรู้</p><p>ไม่มีอะไรภายนอกเปลี่ยนไป</p><p>อาจมีเพียงช่วงหนึ่งบนโต๊ะอาหารที่สายตาของเขากับเธอสบกันนานกว่าปกติเล็กน้อย</p><p>นานพอที่จะจำได้ว่า มีเรื่องหนึ่งที่มีเพียงสองคนบนโต๊ะนั้นเท่านั้นที่รู้</p><p>จากนั้นใครคนหนึ่งก็คงหลบสายตา</p><p>เหมือนที่เคยทำมาตลอด</p><p>ทั้งหมดนั้นเป็นเพียงภาพที่เขาสร้างขึ้นในหัว</p><p>ไม่เคยมีการนัดหมาย</p><p>ไม่เคยมีห้อง</p><p>ไม่เคยมีกล้องตั้งรอเธออยู่ที่ไหน</p><p>ไม่เคยมีการสัมผัสที่ต้องปิดบังในภายหลัง</p><p>และที่สำคัญที่สุด—</p><p>เธอไม่เคยบอกเขาว่า เธอต้องการสิ่งเดียวกัน</p><p>มีเพียงสายตา</p><p>ข้อความที่ตีความได้สองแบบ</p><p>มุกตลกที่ไม่มีใครกล้าถามว่าจริงแค่ไหน</p><p>และช่วงเวลาสั้น ๆ บางช่วงที่ทำให้เขาสงสัยว่า บางทีเรื่องที่เขาไม่กล้าพูดออกมานั้น อาจไม่ได้มีอยู่ในหัวของเขาเพียงคนเดียว</p><p>แต่เขาไม่รู้</p><p>และตราบใดที่ไม่มีใครพูดมันออกมา</p><p>เขาก็ยังไม่จำเป็นต้องรู้</p><p>มันยังคงเป็นเพียง “เรื่องที่พวกเขาไม่ได้รับอนุญาตให้พูดถึง”</p><p>และบางที สิ่งที่ทำให้มันน่าหลงใหลที่สุด</p><p>ก็คือการที่มันยังไม่เคยเกิดขึ้นเลย.​</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=23a8071b4119" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Governance in Maritime]]></title>
            <link>https://medium.com/@inzaghi.phil/ai-governance-in-maritime-7457713a5da8?source=rss-38c1d6aefa3f------2</link>
            <guid isPermaLink="false">https://medium.com/p/7457713a5da8</guid>
            <category><![CDATA[maritime]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[governance]]></category>
            <category><![CDATA[ai-safety]]></category>
            <category><![CDATA[risk-management]]></category>
            <dc:creator><![CDATA[Phillip Inzaghi]]></dc:creator>
            <pubDate>Tue, 11 Aug 2026 15:43:12 GMT</pubDate>
            <atom:updated>2026-08-11T15:43:12.821Z</atom:updated>
            <content:encoded><![CDATA[<p>AI Governance in Maritime</p><p>How to Control AI-Supported Work Through Your Safety Management System</p><p>Why approved tools are not enough—and how verification, document control, competence and accountability apply when AI enters operational work.</p><p>AI is already influencing safety management systems, audit reports and operational decisions inside maritime companies—often without formal approval or governance.</p><p>The question is no longer whether people are using AI. It is whether companies know where, how and under what controls it is being used.</p><p>The risk is not simply that people use AI. The risk is uncontrolled data, unchecked output and misplaced accountability.</p><p>It may be used to draft risk assessments, summarise procedures, prepare toolbox talks, rewrite audit findings, compare requirements, or make safety management documents sound more polished. Some of that use may be helpful. Some of it may be harmless. Some of it may create serious control problems.</p><p>The mistake is treating AI governance as a choice between “ban it” and “allow it”.</p><p>That is too crude.</p><p>The real question is operational: what information is being entered, what output is being produced, who checks it, what evidence supports it, and who remains accountable when AI-supported material enters company use.</p><p>At the time of writing, I have not identified an IMO instrument specifically governing routine employee use of generative AI for tasks such as drafting risk assessments, preparing toolbox talks, reviewing audit findings, summarising regulations, or supporting safety management documentation. IMO has adopted a non-mandatory code for AI-enabled and remotely operated autonomous ships, but that is a different regulatory problem.</p><p>That does not mean companies should wait.</p><p>It means AI use has to be controlled through existing governance duties: safety management, competence, verification, confidentiality, document control, and accountability.</p><p>The First Misunderstanding: AI Is Treated as an IT Issue<br>Many companies will first see AI as an IT or cybersecurity issue.</p><p>That is partly correct, but incomplete.</p><p>Tool access, account type, data retention, permissions, and security settings matter. A company-approved enterprise AI tool is not the same risk as an employee pasting vessel, client, investigation, or personal data into an uncontrolled personal account under time pressure.</p><p>But tool selection does not solve the maritime governance problem.</p><p>A better-controlled tool may reduce some risks. It does not remove the need for review, competence, document control, or operational judgement.</p><p>The company still has to decide:</p><p>what may be entered into AI tools<br>what must never be entered<br>which tools are approved<br>which tasks require verification<br>who is competent to review the output<br>how AI-supported work is recorded where evidence matters<br>Without those controls, AI becomes an informal parallel system. It can influence safety documents, audit records, investigation material, and operational decisions without being visible inside the company’s actual management system.</p><p>Existing Duties Already Cover the Control Problem<br>The ISM Code does not regulate AI.</p><p>But AI-supported work can enter areas the safety management system already controls: procedures, risk controls, competence, records, investigations, audits, review, and company accountability.</p><p>AI Governance Question	Existing SMS Control Area<br>Who remains accountable if AI-supported material is used?	Company responsibility and authority<br>Can this output be used operationally?	Approved procedures and shipboard operational controls<br>Who is competent to check it?	Resources, personnel, competence, and familiarisation<br>Has the source been verified?	Company verification, review, and audit<br>Can the document be changed casually?	SMS documentation and revision control<br>Could this distort a finding, incident, or corrective action?	Reporting and analysis of non-conformities, accidents, and hazardous occurrences<br>Existing maritime regulation was not written for AI.</p><p>However, AI can affect work already controlled by the safety management system. Once AI-supported material enters that system, it still has to be checked, approved, controlled and owned by the company.</p><p>The Second Misunderstanding: Polished Output Is Treated as Reliable Output<br>AI output often sounds confident.</p><p>That is part of the problem.</p><p>In maritime compliance and safety work, a well-written answer is not the same as a verified answer. A risk control can sound sensible and still be unsuitable. A regulatory reference can look plausible and still be wrong. A summary can appear balanced while omitting the controlling requirement.</p><p>Early in my own use of AI, I asked it to help identify an ISM Code reference for a finding. It produced a reference that sounded plausible. When I checked it against the Code, the reference did not exist.</p><p>That was the point where the issue became clear: custom instructions are not controls, and confident output is not evidence.</p><p>The same issue applies to AI-assisted review. A tool can appear to review a document while filling gaps with plausible content that is not actually present in the source.</p><p>In audit, compliance, and safety work, that is not a harmless drafting issue. It can create false findings, false assurance, wasted review time, and misplaced confidence.</p><p>The control is simple in principle, but often missing in practice:</p><p>AI output must be checked against the source before it is relied on.</p><p>Confident output is not evidence.<br>If the output cites a regulation, the regulation must be checked.</p><p>If it summarises a procedure, the procedure must be checked.</p><p>If it proposes a control, the control must be assessed by someone competent to judge whether it is suitable for the operation.</p><p>If it drafts safety management content, the company must still decide whether that content is correct, controlled, and approved.</p><p>The Third Misunderstanding: Accountability Moves to the Tool<br>AI does not take responsibility.</p><p>The person using the output remains responsible for deciding whether it is suitable. The company remains responsible for the management system it operates. A vessel, department, auditor, DPA, superintendent, HSE manager, or master cannot avoid accountability by saying that a tool produced the wording.</p><p>This is where maritime companies should be careful.</p><p>AI can make weak work look complete. It can make uncertain work look confident. It can make generic controls look like a finished risk assessment. It can make an inexperienced user feel as though they have produced something more reliable than they actually have.</p><p>That does not mean AI should never be used.</p><p>It means AI should be treated as an assistant, not an authority.</p><p>What Controlled Use Looks Like<br>A practical maritime AI governance model does not need to begin with a 40-page policy.</p><p>It should begin with clear operating rules.</p><p>At minimum, companies should define:</p><p>approved AI tools and account types<br>prohibited information and upload rules<br>permitted and prohibited task types<br>review requirements for AI-supported outputs<br>source verification requirements<br>document control requirements<br>record retention where AI materially supports a decision or controlled document<br>competence expectations for users and reviewers<br>The strongest control is not the tool itself. It is the workflow around it.</p><p>The strongest control is not the tool itself. It is the workflow around it.<br>A Practical Governance Workflow<br>The workflow can be expressed through four connected controls:</p><p>CONTROL</p><p>Define the approved tool, permitted task and information boundaries.</p><p>VERIFY</p><p>Check claims and proposed controls against the authoritative source and operational context.</p><p>APPROVE</p><p>Require a competent person with the proper authority to accept or reject the output.</p><p>RECORD</p><p>Retain the evidence, review and document history where the output affects a controlled decision or document.</p><p>Human and organisational accountability applies throughout every stage. It does not appear only at the end of the process.</p><p>For example, using AI to improve the readability of a toolbox talk is a different risk from using AI to generate the underlying task risk assessment. Summarising a public guidance note is different from uploading an incident report. Drafting general training prompts is different from asking AI to classify audit findings against a code or regulation.</p><p>Those differences matter.</p><p>A governance system should separate low-risk support tasks from tasks that could affect safety, compliance, legal exposure, client confidentiality, or operational decision-making.</p><p>AI should be treated as an assistant, not an authority.<br>A Simple Control Test<br>Before using AI in maritime QHSE, SMS, audit, or operational assurance work, the user should be able to answer four questions:</p><p>Am I allowed to enter this information into this tool?<br>Is this the right type of tool and account for this task?<br>What source will I check the output against?<br>Who remains responsible if this output is used?<br>If those questions cannot be answered, the task is not controlled.</p><p>The Real Risk Is False Assurance<br>The danger is not that AI makes a spelling mistake.</p><p>The danger is that AI-supported work enters the business looking complete, reviewed, and authoritative when it has not been properly verified.</p><p>That matters in maritime because many documents are not just documents. They are part of a safety management system. They shape work planning, risk control, training, audits, investigations, corrective action, and management review.</p><p>If AI helps draft or review that material, the company needs to know where the output came from, what source supports it, and who checked it.</p><p>Otherwise, the company may create the appearance of control without the substance of control.</p><p>The Operational Dilemma<br>Master reviewing an officer&#39;s hours-of-rest disclosure and compliance records<br>A deck officer hands the Master a written statement.</p><p>For the past two weeks, the officer has been signing rest hour records showing full STCW compliance while actually working beyond legal limits. The vessel is short-handed, the watch schedule is stretched, and the officer has been covering gaps to keep operations moving.</p><p>He is now dangerously fatigued and afraid he will make a navigation error. He comes clean because he would rather face discipline than cause a collision.</p><p>The vessel docks tomorrow for port state control. The rest hour records will be audited. The SMS requires immediate reporting of falsified legal records, disciplinary action, and correction of the official record.</p><p>The Master now has two bad options.</p><p>Option	Immediate Outcome	Governance Risk<br>Correct the records, stand the officer down, and follow the SMS.	The record becomes honest. The vessel may face a clear STCW violation, possible detention, and operational disruption. The officer may be disciplined for admitting the truth.	The crew may learn that honesty destroys the person who reports the problem, making future fatigue reporting less likely.<br>Protect the officer, adjust the watch schedule quietly, and leave the records unchanged.	The vessel may pass inspection. The officer keeps his job and the crew sees that coming forward is protected.	The Master has knowingly retained falsified legal records. If discovered, the issue becomes concealment, not just fatigue management.<br>That is the governance problem.</p><p>A system that only punishes the falsified record may discourage the next exhausted officer from speaking up. But a system that protects honesty by preserving false records destroys the integrity of the SMS and exposes the Master to personal liability.</p><p>Good governance has to deal with both truths at the same time: falsified legal records cannot be ignored, and fatigue reporting cannot be treated as betrayal. If the only available choices are concealment or career destruction, the system has already failed before the officer walks into the Master’s office.</p><p>Key Takeaways<br>Maritime AI governance is a management-system issue, not simply an IT issue.<br>Existing safety management controls already provide much of the necessary governance framework.<br>AI-supported outputs must be verified against authoritative sources and the actual operational context.<br>Accountability remains with the people and organisation using the output.<br>Practical operating controls matter more than lengthy AI policies.<br>Conclusion<br>Maritime companies do not need to wait for AI-specific IMO guidance before acting.</p><p>The basic control problem is already visible.</p><p>AI use should be governed through the same principles that already apply to safety management and assurance: controlled information, competent people, verified sources, approved documents, clear accountability, and evidence where evidence matters.</p><p>The companies that handle this well will not be the ones with the longest AI policy.</p><p>They will be the ones that make AI use visible, controlled, and reviewable before it affects operational work.</p><p>AI will continue to evolve. The principles of good governance will not. Companies that embed AI within their existing management systems, rather than allowing it to operate alongside them as an informal parallel process, will be better placed to use the technology safely, consistently and with confidence.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7457713a5da8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[False Assurance in Maritime Compliance: When Good Records Hide Poor Control
Documentation can show…]]></title>
            <link>https://medium.com/@inzaghi.phil/false-assurance-in-maritime-compliance-when-good-records-hide-poor-control-documentation-can-show-7b2bd87b469c?source=rss-38c1d6aefa3f------2</link>
            <guid isPermaLink="false">https://medium.com/p/7b2bd87b469c</guid>
            <category><![CDATA[maritime]]></category>
            <category><![CDATA[risk-management]]></category>
            <category><![CDATA[safety-management]]></category>
            <category><![CDATA[auditing]]></category>
            <category><![CDATA[compliance]]></category>
            <dc:creator><![CDATA[Phillip Inzaghi]]></dc:creator>
            <pubDate>Tue, 11 Aug 2026 01:59:54 GMT</pubDate>
            <atom:updated>2026-08-11T01:59:54.917Z</atom:updated>
            <content:encoded><![CDATA[<h3>False Assurance in Maritime Compliance: When Good Records Hide Poor Control<br>Documentation can show that a control exists. Only operational evidence can show whether it works.<br>By Phillip Inzaghi — Founder, AxiomOrdo; Maritime QHSE Specialist<br>Maritime companies rarely lack paperwork.<br>They have policies, procedures, certificates, permits, audit questions, training records, performance dashboards and corrective-action systems.<br>These records are necessary. They establish expectations, allocate responsibility and provide evidence that required activities have taken place.<br>The problem begins when their existence is treated as proof that the underlying risk is controlled.<br>A procedure describes how work should be done. A permit records the conditions under which work has been authorised. A certificate shows that a ship or management system has been inspected against a defined scope.<br>None of these, by itself, proves that the control is understood, implemented or effective during real operations.<br>That gap between documented compliance and operational reality is false assurance.<br>What false assurance means<br>Assurance is justified confidence that controls are appropriately designed, consistently implemented and effective in practice.<br>False assurance occurs when an organisation mistakes evidence that a control exists for evidence that the control works.<br>Typical statements include:<br>We have a policy for that.<br>It is covered by the safety management system.<br>The permit was signed.<br>The training was completed.<br>The audit found no major non-conformities.<br>The ship holds the required certificate.<br>The dashboard is green.<br>Each statement may be factually correct.<br>The mistake is assuming that it ends the enquiry.<br>The four alignments of credible assurance<br>Credible assurance requires alignment across four areas.<br>1. Requirement<br>What does the applicable regulation, standard or company requirement actually require?<br>2. Documentation<br>Has that requirement been translated accurately into the company’s system?<br>3. Understanding<br>Do affected personnel understand the control and how to apply it?<br>4. Practice<br>Is the control implemented and effective during real work?<br>When these elements do not align, apparently strong records can conceal weak control.<br>How an assurance system should validate control<br>A sound assurance process should trace the requirement through the entire system:<br>Requirement → Documentation → Understanding → Practice → Validation<br>The requirement must first be understood correctly.<br>It must then be translated into appropriate controls and accurately captured in policies, procedures and work instructions.<br>Affected personnel must understand those controls.<br>Finally, implementation and effectiveness must be verified in operational practice.<br>Only then is the assurance chain complete.<br>When an incomplete system audits itself<br>I encountered a clear example while reviewing a company’s arrangements under the Maritime Labour Convention, 2006.<br>The company’s safety management documentation ran to approximately 98 pages, but its coverage of the MLC was limited to roughly half a page.<br>Its internal audit framework contained 16 MLC-related questions. When those questions were compared with the SMS, only six could be traced to identifiable supporting provisions.<br>The company could point to an MLC section in the SMS and an MLC audit checklist and conclude that the subject was covered.<br>The difficulty was that both the documented framework and the audit criteria were incomplete.<br>The result was a closed assurance loop:<br>the SMS defines too little;<br>the internal audit tests only what the company has chosen to define;<br>satisfactory answers are treated as evidence of compliance;<br>management receives confidence from a framework that has not tested the full requirement.<br>In other words:<br>Control gap → incomplete SMS coverage → incomplete audit criteria → management confidence → false assurance<br>Certification provides legitimate evidence, but the MLC itself does not treat a certificate as an unlimited guarantee.<br>Regulation 5.1.1 describes the Maritime Labour Certificate and DMLC as prima facie evidence that the ship has been inspected and that requirements have been met &quot;to the extent so certified.&quot;<br>The same regulation requires an effective system ensuring that working and living conditions &quot;meet, and continue to meet&quot; the Convention’s standards.<br>Certification is evidence within a defined scope.<br>It does not remove the need for continuing verification.<br>The certificate existed, but the requirement was not met<br>On every vessel I personally visited during this review, the Maritime Labour Certificate and DMLC were onboard.<br>When I asked where the crew could see them, the answer was that the documents were retained in the Master’s certificate folder.<br>The people involved were not deliberately concealing them.<br>They treated the documents like other statutory certificates: valid, onboard and available if an inspector requested them.<br>Standard A5.1.3 is more specific.<br>It requires the current Maritime Labour Certificate and DMLC to be carried onboard and copies to be posted in a conspicuous place where they are available to seafarers.<br>The documents existed, but the purpose of the requirement had not been achieved.<br>The issue was corrected on one vessel, but the same condition was then found on another.<br>That showed that the first response had addressed a local condition rather than the management system.<br>A practical fleetwide response would have been straightforward: define a consistent conspicuous location and incorporate it into the fleet’s documented arrangements and verification process.<br>The visible problem had been corrected. The system that allowed it had not.<br>Internal administrative practice is not regulatory authority<br>The same review identified gaps involving seafarers&#39; employment agreements, collective bargaining arrangements, contractual terms and wage information.<br>Regulation 2.1 and Standard A2.1 require clear, written and legally enforceable employment terms, signed originals for the parties, employment information available onboard, applicable collective bargaining agreements onboard, and wage particulars in the agreement.<br>Following a company-wide pay rise, updated contracts had not been issued.<br>HR instead placed a note in its system stating that contractual pay had been updated.<br>Whether that method was legally sufficient would depend on the applicable flag-State law, agreement terms and collective bargaining arrangements.<br>The assurance failure was more basic: an internal administrative method had been adopted without demonstrating how it satisfied the applicable external requirements.<br>A payroll system may prove that the correct amount was paid.<br>An HR note may prove that the organisation recorded a change.<br>Neither automatically proves that the applicable regulatory and contractual requirements have been met.<br>A permit does not make work safe<br>A permit is required because the proposed task presents higher or less routine risk and therefore needs specified conditions before work begins.<br>The permit itself does not make the job safer.<br>Protection comes from the isolations, tests, barriers, competent personnel, communications, rescue arrangements, equipment and worksite readiness behind it.<br>False assurance arises when the completed document is treated as proof that those controls exist.<br>For familiar work, a permit issuer may believe that they already know the area and sign without visiting the worksite.<br>The greater danger may be the condition that is different today:<br>another task nearby;<br>changed access;<br>moved equipment;<br>a missing barrier;<br>a live energy source;<br>or a work area that no longer matches the plan.<br>A permit based on yesterday’s knowledge cannot confirm today’s conditions.<br>Requiring an inspection in a procedure is not enough.<br>The organisation must establish how it will ensure that the inspection actually occurs.<br>Each control needs a distinct purpose<br>A strong work-control system does not simply accumulate documents. Each element should have a clear function.<br>Risk assessment<br>Identifies hazards, evaluates risk and defines controls.<br>It should not become a full repetition of the job plan.<br>Job plan or method<br>Describes sequence, responsibilities, critical stages, resources, hold points and contingency arrangements.<br>Permit to work<br>Authorises a defined higher-risk activity within stated limits and records the conditions required before work begins.<br>Pre-job briefing<br>Confirms understanding, focuses attention on critical stages and invites operational input.<br>It should not consist of one person reading every possible hazard while the team disengages.<br>Start Work Verification Check<br>Provides final physical verification at the worksite.<br>An independent person asks for controls to be demonstrated:<br>Show me the isolation.<br>Show me the barrier.<br>Show me the required equipment.<br>What has changed?<br>Are all permit conditions actually in place?<br>The permit records what should exist.<br>The Start Work Verification Check confirms that it does.<br>How auditors test whether the system works<br>A good system can be explained by the people operating it.<br>Valuable evidence often comes from asking personnel to describe their work rather than testing whether they can repeat a procedure word for word.<br>The discussion can begin simply:<br>Tell me about your role onboard and the work you normally perform.<br>From there, ask:<br>Where is the documentation?<br>What is each document for?<br>How does planning occur?<br>Do you contribute to the briefing?<br>What happens if conditions change?<br>When would you stop the job?<br>How is the task closed?<br>A weak audit reads each checklist question and records the answer received.<br>A stronger audit uses the checklist to confirm coverage but follows the evidence through conversation, documents, records and observation.<br>The checklist supports the auditor. It should not replace judgement.<br>When actual practice differs from the procedure<br>A difference between written procedure and actual practice is not automatically proof that the worker is wrong.<br>It is a reason to investigate.<br>The person may:<br>not know the procedure;<br>be using an unsupported shortcut;<br>be compensating for an impractical instruction;<br>have developed a safer method;<br>or not know how to request a document change.<br>A fleet should operate as a learning system.<br>But the existence of a document-change system does not prove that learning occurs.<br>If a work instruction has accumulated 20 unresolved requests, the organisation already has evidence that it may be outdated, unclear or impractical.<br>Where resources are constrained, the backlog must be prioritised by risk.<br>A minor editorial amendment can wait.<br>A work instruction capable of contributing to injury, non-conformity or detention cannot.<br>How false assurance moves upward<br>Senior leaders often genuinely want the system to work properly.<br>The difficulty may arise in middle management, where escalating a weakness means asking for money, people, time, operational disruption or uncomfortable scrutiny.<br>Issues may therefore be softened, narrowed or treated as isolated.<br>Management review should distinguish between one-off deficiencies and recurring patterns across vessels or activities.<br>Senior management should receive a clear decision package:<br>what happened;<br>why it happened;<br>what has already been tried;<br>why that was insufficient;<br>what must happen next;<br>how urgent it is;<br>the consequences of delay;<br>realistic solution options;<br>and the financial and resource implications.<br>Assurance should support decisions, not simply produce reassurance.<br>Closing an action is not proving effectiveness<br>Systemic issues should not be closed because management has agreed to update a policy.<br>A revised document proves that an action was taken.<br>It does not prove that the problem was solved.<br>Effective closure means demonstrating that:<br>the control was revised;<br>related documents and forms were updated;<br>the change was communicated;<br>affected personnel understood it;<br>the change was implemented onboard;<br>observed practice changed;<br>the intended outcome improved over time; and<br>the original failure stopped recurring.<br>Until effectiveness has been demonstrated, the organisation has changed its paperwork, not necessarily its control.<br>Five warning signs of false assurance<br>1. A self-audit reports no findings<br>Repeated clean audits in a complex operational system should trigger scrutiny of sampling, depth, interviews, observation and willingness to challenge.<br>2. Controls are documented or certified but not validated in operation<br>Policies, procedures, permits and certificates exist, but no one tests whether controls work under actual conditions.<br>3. Assurance activity reduces anxiety rather than risk<br>Forms, signatures, meetings and approvals are added because they make the organisation feel protected, without materially improving control.<br>4. Indicators remain green because they are lagging or aggregated<br>Examples include:<br>high audit pass rates where testing was limited;<br>no lost-time injuries while near misses or under-reporting rise;<br>100% training completion without competence being tested;<br>corrective actions closed administratively without effectiveness review.<br>5. Crew cannot locate or explain task-specific documentation<br>If personnel cannot find the relevant document, explain what each stage is for or describe what to do when conditions change, the management system is not functioning as intended.<br>Operational dilemma: the green fleet dashboard<br>Imagine you are the Designated Person Ashore for a company operating 18 vessels.<br>The management-review dashboard shows:<br>97% audit compliance;<br>no major non-conformities;<br>the lowest lost-time injury frequency in five years;<br>100% training completion;<br>96% permit compliance;<br>94% corrective actions closed on time.<br>Everything appears positive.<br>Then several signals emerge:<br>audits relied heavily on records with limited observation;<br>permit issuers signed site checks without visiting;<br>crew could not distinguish the controls;<br>high-risk procedures had unresolved change requests;<br>an MLC deficiency recurred after closure.<br>You now have two choices.<br>Choice A: preserve confidence while investigating quietly<br>Continue presenting the dashboard and commission a limited review framed as continuous improvement.<br>This protects short-term confidence but risks later appearing to conceal known warning signs.<br>Choice B: escalate the uncertainty immediately<br>Tell senior management that the indicators may not accurately represent operational control, identify what remains uncertain and commission a risk-based independent review.<br>I would choose Choice B.<br>That does not require declaring the whole SMS ineffective.<br>It requires reporting that several independent signals now call the reliability of the assurance indicators into question.<br>The numbers may still be correct.<br>The conclusion drawn from them may no longer be.<br>AI can make false assurance more convincing<br>AI does not create false assurance, but it can make it easier to produce and harder to recognise.<br>It can rapidly generate:<br>procedures;<br>risk assessments;<br>audit checklists;<br>corrective-action plans;<br>training material;<br>management summaries.<br>Polished output may appear more complete, current and authoritative than the underlying evidence supports.<br>AI-generated material can reproduce earlier assumptions, copy incomplete controls across multiple documents or introduce plausible but incorrect regulatory statements.<br>AI can support regulatory comparison, gap identification, document review and preparation.<br>It should not convert unverified information into a final compliance conclusion.<br>Claims still need checking against primary sources.<br>Applicability must be confirmed for the vessel and flag.<br>Generated documents must be compared with actual work.<br>The more convincing the documentation becomes, the more important operational verification becomes.<br>A practical assurance test<br>A useful assurance review can be built around six stages.<br>1. Requirement<br>What requirement is being addressed?<br>Has it been fully mapped?<br>Is the interpretation current and applicable?<br>2. Documentation<br>Is the control accurately documented?<br>Is it current and consistent across procedures, forms and instructions?<br>3. Understanding<br>Can affected personnel find the relevant documentation?<br>Can they explain its purpose?<br>Can they describe what to do when conditions change?<br>4. Implementation<br>Is the stated control visible in practice?<br>Has the activity been observed?<br>Do the records reflect what actually occurred?<br>5. Effectiveness<br>Has the control reduced the intended risk?<br>Are findings recurring?<br>Have outcomes improved?<br>6. Learning<br>Are deviations investigated?<br>Are changes processed?<br>Are improvements shared?<br>Are recurring issues escalated?<br>A completed document may answer an administrative question. Only operational evidence can show whether the organisation is actually in control.<br>Conclusion<br>Maritime assurance is not created when paperwork looks correct.<br>It is created when applicable requirements are translated into workable controls, personnel understand why those controls exist, actual conditions are verified and management responds when evidence shows that the system is not working.<br>Policies, permits, certificates, audits and dashboards remain essential.<br>Their value lies in what they reliably demonstrate, not in their existence.<br>Compliance is demonstrated not when the paperwork is complete, but when the intended control consistently works under real operating conditions.<br>The final question is not:<br>Is the document complete?<br>It is:<br>Can the organisation demonstrate that the work, obligation or risk is actually under control?<br>References<br>Maritime Labour Convention, 2006 — Regulation 2.1 and Standard A2.1: Seafarers&#39; employment agreements.<br>Maritime Labour Convention, 2006 — Regulation 5.1.1: General principles of flag-State inspection and certification.<br>Maritime Labour Convention, 2006 — Regulation 5.1.3 and Standard A5.1.3: Maritime Labour Certificate and Declaration of Maritime Labour Compliance.<br>Phillip Inzaghi<br>Founder, AxiomOrdo<br>Maritime QHSE Specialist<br>phil@axiomordo.com</h3><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7b2bd87b469c" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>