<?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 Praful S. on Medium]]></title>
        <description><![CDATA[Stories by Praful S. on Medium]]></description>
        <link>https://medium.com/@praful-shah?source=rss-113fd96adab9------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Stories by Praful S. on Medium</title>
            <link>https://medium.com/@praful-shah?source=rss-113fd96adab9------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 09:09:39 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@praful-shah/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[SF4COE : Three Roads to OmniStudio Standard Runtime: Hybrid, Migration, and Future-Ready]]></title>
            <link>https://medium.com/@praful-shah/sf4coe-three-roads-to-omnistudio-standard-runtime-hybrid-migration-and-future-ready-55a3c241ef4a?source=rss-113fd96adab9------2</link>
            <guid isPermaLink="false">https://medium.com/p/55a3c241ef4a</guid>
            <dc:creator><![CDATA[Praful S.]]></dc:creator>
            <pubDate>Wed, 10 Jun 2026 20:57:46 GMT</pubDate>
            <atom:updated>2026-06-10T20:57:46.582Z</atom:updated>
            <content:encoded><![CDATA[<p><em>How to choose the right adoption strategy for your org — and why doing nothing is the only wrong answer.</em> — —</p><p>If you run Vlocity Industries (Communications, Insurance, Health, Energy &amp; Utilities, or Public Sector), you’ve probably heard the message from Salesforce by now: the <strong>managed package runtime is the past, and the Standard Runtime is the future</strong>. For years, OmniScripts, FlexCards, Integration Procedures, and DataRaptors lived inside custom objects like vlocity_cmt__OmniScript__c, executed by a managed package engine, and deployed with the Vlocity Build Tool. The Standard Runtime changes all of that. Components now live on standard Salesforce objects, run on a platform-native execution model, deploy through the Salesforce CLI, and benefit from faster feature delivery and first-class Metadata API support. But here&#39;s the part many teams underestimate: <strong>this is not an upgrade you toggle on. It&#39;s a runtime and data model migration.</strong> Nothing gets refactored automatically, not every component is compatible, and once you flip the switch in production, rolling back is painful. That reality leaves architects with a strategic question rather than a technical one: <em>how do we get from here to there without breaking the business?</em> In my experience, every successful adoption story falls into one of three approaches. Let&#39;s walk through them. - -</p><h3>Approach 1: The Hybrid Approach — Run Both Runtimes Side by Side</h3><p>Switch Between Standard and Managed Package Runtimes and Designers in the Same Org Omnistudio Hybrid</p><p><strong>Best for:</strong> Large, complex orgs with hundreds of OmniStudio components, heavy customization, and zero tolerance for downtime. The hybrid approach accepts a simple truth: you cannot migrate everything at once, and you shouldn’t try. Instead, you operate the managed package runtime and the Standard Runtime in parallel for a defined transition period.</p><h3>How it works</h3><ul><li><strong>Net-new development happens on Standard Runtime.</strong> Every new OmniScript, FlexCard, or Integration Procedure is built natively from day one. You stop adding to the legacy pile.</li><li><strong>Existing components stay on the managed runtime</strong> until they’re scheduled for migration in prioritized waves — typically grouped by business capability (quoting journeys first, service journeys second, and so on).</li><li><strong>A migration backlog governs the transition.</strong> Each wave goes through assess → remediate → migrate → regression test, just like a mini-project.</li></ul><p>Why teams choose it</p><ul><li>The hybrid model de-risks the transition. Revenue-critical journeys keep running on the runtime they were tested against, while the team builds muscle memory on the new runtime with lower-stakes work. It also spreads the cost: instead of one massive migration project competing with your roadmap, migration becomes a steady workstream alongside feature delivery.</li></ul><p>The trade-offs</p><ul><li>Running two runtimes means two deployment pipelines, two sets of patterns, and two mental models for your developers. Documentation, training, and governance overhead are real. And there’s a gravitational pull toward procrastination — without a hard sunset date for the managed runtime, “temporary” hybrid states have a way of becoming permanent. Set a deadline and defend it. —</li></ul><h3>Approach 2: The Migration Approach — A Structured, Full Cutover</h3><ul><li><strong>Best for:</strong> Mid-sized orgs with a manageable component inventory, or any team that wants to rip the bandage off and consolidate quickly. This is the classic, project-based migration — and Salesforce gives you purpose-built tooling for it: the <strong>OmniStudio Migration Assistant (OMA)</strong>, a Salesforce CLI plugin with two execution modes.</li><li># Install the plugin sf plugins install @salesforce/plugin-omnistudio-migration-tool # Assess your org&#39;s readiness sf omnistudio:migration:assess -u your-org-alias # Run the migration sf omnistudio:migration:migrate -u your-org-alias -a</li><li><strong>Assess mode</strong> scans the org, flags unsupported components, and produces a readiness report. <strong>Migrate mode</strong> converts eligible components to their standard-runtime equivalents and logs everything. The golden rule: <em>never run migrate until your assessment report is clean or every flagged item is understood.</em></li></ul><p>The four phases that make it work</p><p><strong>Phase 0 — Preparation.</strong> Inventory everything: OmniScripts, Integration Procedures, DataRaptors, Calculation Procedures, custom LWCs, and the Apex behind them. Enable the OmniStudio Metadata API. Pick a development sandbox as your first target. Teams that skip this phase are the ones whose migrations fail.</p><p><strong>Phase 1 — Sandbox migration.</strong> Run assess, fix what it finds, re-run until acceptable, then migrate. Expect three buckets of outcomes: components that convert cleanly, components needing manual refactoring, and components that are simply unsupported.</p><p><strong>Phase 2 — Testing and validation.</strong> Deploy to QA/UAT and run full regression: OmniScript navigation and data handling, Integration Procedure inputs and outputs, FlexCard rendering, Experience Cloud pages. This is where subtle behavioral differences surface — error handling and UI rendering are common culprits, and even small oddities appear (for example, DataRaptor Transforms between JSON and XML occasionally migrating with the wrong output type, which you fix by editing the migrated copy).</p><p><strong>Phase 3 — Production rollout.</strong> If the earlier phases were done right, production should be boring. Deploy via CLI or CI/CD, validate key user journeys, disable the managed runtime, enable Standard Runtime settings. Then hold your breath — there’s no easy rollback.</p><p>Know the manual work up front</p><ul><li>Automation covers a lot 70–80%, but not everything:</li><li><strong>Apex remote actions</strong> implementing vlocity_cmt.VlocityOpenInterface must be refactored to implement System.Callable.</li></ul><p><strong>Custom LWCs embedded in OmniScripts</strong> need updating to the standard-runtime base components, and remote calls to Integration Procedures or DataRaptors must follow the new OmniScript action framework (don’t forget to enable <em>Deploy Custom Lightning Web Components in Standard Runtime</em> in Setup, and add &quot;useCustomLMS&quot;: true to the OmniScript JSON if your LWC uses messaging).</p><ul><li><strong>Legacy Angular OmniScripts</strong> must be rebuilt entirely.</li><li><strong>Calculation Procedures</strong> often need to move to the Business Rules Engine.</li><li><strong>Reserved variable names</strong> like request and response can silently break Integration Procedures.</li></ul><p>The trade-offs</p><ul><li>A full cutover concentrates risk into a single window and demands disciplined regression testing. It also temporarily freezes feature delivery — your roadmap takes a back seat while the migration runs. But you come out the other side with one runtime, one toolchain, and no lingering technical debt. — -</li></ul><p>Approach 3: The Future-Ready Approach — Standard Runtime First</p><ul><li><strong>Best for:</strong> New implementations, orgs early in their Industries journey, or teams using the migration mandate as a trigger for a broader redesign. The third approach reframes the question. Instead of asking <em>“how do we move what we have?”</em>, it asks <em>“if we were building this today, what would it look like?”</em></li></ul><p>How it works</p><ul><li><strong>Greenfield builds start on Standard Runtime — no exceptions.</strong> New orgs and new business units never touch the managed package runtime.</li><li><strong>Existing assets are selectively rebuilt, not migrated.</strong> Rather than converting a five-year-old OmniScript with 40 steps and a tangle of conditional logic, you redesign the journey using current best practices — Business Rules Engine instead of Calculation Procedures, modern LWC patterns, leaner Integration Procedures.</li><li><strong>DevOps is modernized from day one.</strong> Source-driven development, Salesforce CLI deployments, CI/CD pipelines, and standard metadata management replace legacy Vlocity build processes entirely.</li><li><strong>Anything not worth rebuilding gets retired.</strong> A migration mandate is the best inventory audit you’ll ever get. Most mature orgs discover a meaningful percentage of their components are unused or redundant.</li></ul><h3>Why teams choose it</h3><p>Migration tools preserve your design decisions — including the bad ones. The future-ready approach treats the runtime transition as a platform modernization win rather than a compliance exercise. You shed years of accumulated workarounds, align fully with Salesforce’s strategic direction, and position the team to adopt new OmniStudio features the moment they ship — because that’s where Salesforce is investing.</p><h3>The trade-offs</h3><p>Rebuilding costs more up front than converting, and it requires strong product ownership to re-validate requirements that may not have been documented since the original implementation. It’s a strategy for teams with budget, executive sponsorship, and a genuine appetite for change — not a shortcut. — -</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/655/1*Ro86EthhaYufYWWOZT-LIA.png" /><figcaption>Choosing Your Road</figcaption></figure><h3>Future Ready — New Org Approach.</h3><ol><li><strong>If your assessment report is mostly green</strong>, the straight migration approach is the fastest path to value.</li><li><strong>If your org is large, customized, and revenue-critical</strong>, hybrid buys you safety — but commit to a sunset date in writing.</li><li><strong>If your implementation is old enough that nobody remembers why half the OmniScripts exist</strong>, future-ready will cost more and be worth every dollar. And in practice? Most enterprises blend them: hybrid governance at the program level, OMA-driven migration for healthy components, and selective rebuilds where the design has aged badly. — -</li></ol><p>The Bottom Line</p><p>The move to OmniStudio Standard Runtime is inevitable — Salesforce has made the strategic direction unambiguous, and staying on the managed runtime means falling behind on features, tooling, and support. The only real decision is <em>how</em> you make the journey. Whichever road you take: <strong>assess before you migrate, test like the business depends on it (it does), and treat this as the modernization opportunity it actually is.</strong> Done right, you don’t just end up compliant — you end up with a faster, cleaner, more future-proof OmniStudio implementation. — — <em>Have you started your Standard Runtime journey? I’d love to hear which approach your team chose and what surprised you along the way — drop a comment below.</em></p><p><strong>Tags:</strong> #Salesforce #OmniStudio #Vlocity #SalesforceIndustries #Migration #Architecture</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=55a3c241ef4a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[When DevOps met Agile at scale: implementation stories and the outcomes that changed how we work]]></title>
            <link>https://medium.com/@praful-shah/when-devops-met-agile-at-scale-implementation-stories-and-the-outcomes-that-changed-how-we-work-7d5edb952309?source=rss-113fd96adab9------2</link>
            <guid isPermaLink="false">https://medium.com/p/7d5edb952309</guid>
            <dc:creator><![CDATA[Praful S.]]></dc:creator>
            <pubDate>Fri, 27 Mar 2026 02:57:49 GMT</pubDate>
            <atom:updated>2026-03-27T02:57:49.882Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Real enterprise stories, hard-won lessons, and the patterns that actually move the DORA metrics needle when you’re operating across 20+ teams.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*d5OTvSeI2znTyWF_Bnr-Jw.png" /></figure><p><em>DevOps is not a team. Agile is not a ceremony. Together, they are a delivery system — and systems need to be designed, not just adopted</em></p><h3>Story 1 — The global bank that shipped every eight weeks</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/344/1*C51E6HvvxMvJmBByUSxvtA.png" /></figure><p>The starting point</p><p>Eight-week release cycles. Three change approval boards. A deployment process that required a 47-step manual runbook. The team had adopted Agile ceremonies faithfully — standups, retrospectives, PI Planning — but the software still sat in a testing queue for weeks before it ever reached a user.</p><h3>What changed</h3><p>The breakthrough wasn’t a tool. It was the realisation that the CAB process and the Agile sprint were completely decoupled. You could complete ten sprints of perfect work and still wait six weeks for a change window. The team rebuilt the change approval process as a set of automated gates — the same logic that had lived in the CAB checklist now ran as pipeline checks in the CI/CD system. What had been a three-day human review became a 22-minute automated verification.</p><p>They also adopted versioned DevStream branches: devstream/3.1.100 for the core banking team, devstream/3.1.200 for the digital channels team. Both streams converged at release/3.1.X with a defined gate that mirrored the old CAB criteria. The Release Manager still approved the final promotion to production — but only after the system had already verified everything the CAB used to verify manually.</p><p><strong>Outcome after 6 months:</strong> Deployment frequency went from every 8 weeks to 3 times per week. Lead time dropped from 47 days to 4 days. Change failure rate fell from 18% to 6%. The CAB still exists — but it now meets for exceptions only, not routine releases.</p><h3>Story 2 — The SaaS scale-up drowning in merge conflicts</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/283/1*yHNWTn02bU4_kOCjdIZ9HA.png" /></figure><p>The starting point</p><p>A shared develop branch serving 14 feature teams. Every sprint end looked the same: a 72-hour merge freeze while the release manager tried to reconcile two weeks of parallel development into a single deployable artifact. Engineers dreaded Fridays. Product managers stopped promising dates. The velocity metrics were fine; the throughput was terrible.</p><h3>What changed</h3><p>The team introduced versioned DevStream branches with a strict naming convention: MAJOR.MINOR.TEAM_ID. Each squad got its own stream — devstream/4.2.100, devstream/4.2.200, devstream/4.2.300 — with a lifecycle tied to the release quarter. Streams synced from develop weekly. They merged into release/4.2.X only when ready, with a defined integration test suite that ran automatically on each PR.</p><p>The key cultural shift: teams were now responsible for the health of their own stream. A broken stream didn’t cascade to other teams. The merge freeze disappeared because the hard integration work happened continuously throughout the sprint, not in a 72-hour panic at the end.</p><p><strong>Outcome after one quarter:</strong> Merge conflicts reduced by 80%. Sprint-end merge freeze eliminated entirely. MTTR fell from 4 hours to 22 minutes because rollback now targeted a single stream, not the entire release. Developer satisfaction (eNPS) rose 34 points.</p><h3>Story 3 — The healthcare platform that couldn’t afford a production incident</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/386/1*Dtr1lFBA67rZ7odZMNbOMg.png" /></figure><h3>The starting point</h3><p>A patient-facing platform operating under HIPAA and FDA software regulations. Any production incident triggered a mandatory incident report and a 30-day remediation audit. The team deployed infrequently — once per month, sometimes less — not because they couldn’t move faster, but because the risk of a failed deployment felt existential. The irony: infrequent deployments meant each release was large, risky, and more likely to fail.</p><h3>What changed</h3><p>They implemented a four-gate environment pipeline: DEV → QA → STG → PROD. Each gate had a defined test suite, a named approver, and a documented artefact proving what was verified. The key insight was that this structure was exactly what the regulatory framework required — they had always needed to prove these checks happened, but previously the proof was scattered across emails, tickets, and hallway conversations.</p><p>By formalising the gate as code — automated tests, structured approvals, immutable audit logs — they simultaneously satisfied their compliance team and their engineering team. The compliance officer became one of the strongest advocates for the pipeline because it generated the evidence trail they had always been trying to maintain manually.</p><p><strong>Outcome after 3 sprints:</strong> Zero production incidents across 47 deployments. Deployment frequency increased from monthly to weekly. Audit preparation time for each release dropped from 3 days to 4 hours. Regulatory review pass rate: 100%.</p><h3>The starting point</h3><p>A company running SAFe with OKRs, but the two systems weren’t talking to each other. Teams would meet their sprint commitments and still fail their Key Results because the features they shipped weren’t the ones that moved the business metrics. There was no explicit link between the work on the board and the outcome on the OKR dashboard.</p><h3>Story 4 — The retail enterprise aligning OKRs to deployments</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/336/1*4A2vte255Y_CL78cUnNwWg.png" /></figure><h3>What changed</h3><p>The team introduced feature flags as the connective tissue between engineering delivery and business outcomes. A feature could be shipped to production — fully tested, behind a flag — before the business was ready to turn it on. This separated technical delivery from business activation. Sprint goals became “ship the flag” rather than “launch the feature.” OKR measurement happened at flag activation, not at code merge.</p><p>Combined with a DevStream model where each product team owned their stream, they could now ship multiple flag-controlled features per sprint, activate them on a business schedule, and roll back any individual flag without a deployment. The NPS improvement came from being able to A/B test activations — turning the feature on for 10% of users, measuring impact, then expanding confidence.</p><p><strong>Outcome after two PI cycles:</strong> OKR hit rate improved from 44% to 71%. NPS lifted by 22 points, attributed directly to the A/B activation of three features. Rollback incidents fell to zero — because “rollback” now meant toggling a flag, not reverting a deployment.</p><h3>The patterns that appear in every successful implementation</h3><h3>1. Automate the human bottlenecks, not just the technical ones</h3><p>Every organisation above had human approval steps that were genuinely important — regulatory compliance, risk management, business sign-off. The mistake is thinking DevOps means removing those humans. The right move is automating the preparation work so that when a human does review something, the review is meaningful. A Release Manager who only sees green pipelines and structured evidence approves in minutes. One who must manually verify a 47-step checklist takes days — and misses things.</p><p><strong>Pattern:</strong> Map every human approval to the specific information it requires. Then automate the collection of that information. The human approves; the system provides the proof.</p><h3>2. Branch names are release plans — version them accordingly</h3><p>The naming convention devstream/MAJOR.MINOR.TEAM_ID appears trivial until you&#39;re in a production incident at 2am trying to determine which stream introduced the breaking change. Branch names that carry semantic meaning — team, release cycle, version — are self-documenting. They eliminate an entire category of &quot;where did this come from?&quot; questions.</p><h3>3. The sprint boundary and the deployment boundary are different things</h3><p>Agile sprints define a planning rhythm. Deployments define a delivery rhythm. Conflating the two — assuming that every sprint should end with a deployment — creates artificial pressure that leads to rushed releases and Friday afternoon outages. Feature flags decouple these rhythms. A two-week sprint can produce ten deployments, or one, depending on what the business needs. What matters is that the code is always releasable.</p><h3>4. Measure DORA from day one, even before you’re proud of the numbers</h3><p>Every team above started measuring DORA metrics before they were performing well. This is deliberate. You cannot improve what you do not measure, and the discomfort of seeing a 47-day lead time on a dashboard is a more powerful forcing function than any consultant’s slide deck. Make the numbers visible. Make them honest. Make them part of every PI retrospective.</p><h3>5. Team topology determines your DevStream topology</h3><p>The most common failure mode in versioned DevStream implementations is mapping streams to features rather than to teams. Features end and streams become orphaned. Teams are more durable. A DevStream should map to a team’s scope of ownership — and when a team is restructured, the stream changes with it. The stream is an expression of team topology, not project structure.</p><h3>What the data says across all four stories</h3><p>4×</p><p>Avg deploy frequency increase</p><p>↓67%</p><p>Avg lead time reduction</p><p>↓58%</p><p>Avg failure rate reduction</p><p>↑41pt</p><p>Avg developer NPS gain</p><p>None of these results came from a tool purchase. They came from deliberate system design — treating the combination of Agile practices, DevOps tooling, and team structure as a single integrated delivery system. The organisations that achieved these results didn’t ask “should we do Agile or DevOps?” They asked “how do these practices reinforce each other in our specific context?” That question leads somewhere useful.</p><blockquote>“The pipeline is not the transformation. The pipeline is the evidence that the transformation happened.”</blockquote><h3>Where to start</h3><p>If your organisation is earlier in this journey, the sequence that appears most reliably in the literature and in these stories is:</p><ul><li>Measure your current DORA state — honestly, even if the numbers are painful</li><li>Identify your longest human wait step in the delivery pipeline — that is your first automation target</li><li>Introduce versioned branch naming before you introduce new tooling — naming is free and immediately reveals structural problems</li><li>Implement the DEV → QA → STG → PROD gate pipeline with explicit owners at each stage</li><li>Adopt feature flags for any feature that has a business activation date — this one change unlocks the decoupling of technical and business delivery rhythms</li><li>Revisit DORA metrics quarterly and treat regression as a team-level incident</li></ul><p>The transformation is not a project. It is a system that you operate, measure, and improve continuously — which, if you have been doing Agile correctly, is exactly what you have been practicing all along.</p><p>#DevOps#Agile#SAFe#DORA#CICD#ReleaseEngineering#PlatformEngineering#TeamTopology#FeatureFlags#EngineeringLeadership</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7d5edb952309" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Vibe Coding: A Step-by-Step Approach to AI-Driven Software Development]]></title>
            <link>https://medium.com/@praful-shah/vibe-coding-a-step-by-step-approach-to-ai-driven-software-development-5749c163fbe6?source=rss-113fd96adab9------2</link>
            <guid isPermaLink="false">https://medium.com/p/5749c163fbe6</guid>
            <dc:creator><![CDATA[Praful S.]]></dc:creator>
            <pubDate>Mon, 24 Mar 2025 01:54:54 GMT</pubDate>
            <atom:updated>2025-03-24T01:54:54.526Z</atom:updated>
            <content:encoded><![CDATA[<h4>Introduction to Vibe Coding</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VlIuvT6a6gQ7jVlo0CU_Ww.jpeg" /></figure><p>Vibe coding is a modern software development methodology that integrates artificial intelligence (AI) into the coding process. It enables developers to focus on strategic decision-making by automating repetitive tasks like code generation, debugging, and optimization.</p><h3>Step-by-Step Guide to Vibe Coding</h3><h3>Step 1: Set Up the Development Environment</h3><p>Before you start with vibe coding, you need the right tools and an AI-integrated development environment (IDE).</p><p>🔹 <strong>Choose an AI-supported IDE:</strong> Popular options include <strong>Cursor</strong>, <strong>Windsurf</strong>, or <strong>GitHub Copilot</strong>, which seamlessly integrate AI models.<br> 🔹 <strong>Select an AI Model:</strong> Claude 3.7, OpenAI’s GPT-4, and Grok are commonly used for vibe coding.<br> 🔹 <strong>Install Required Extensions:</strong> Many IDEs have AI-powered plugins that assist in code generation and debugging.</p><p><strong>Example:</strong> If using Cursor, you can integrate GPT-4 Turbo for smart code suggestions and refactoring.</p><h3>Step 2: Define Project Requirements &amp; Objectives</h3><p>To ensure AI understands the coding task, clear specifications are crucial.</p><p>🔹 <strong>Break Down the Project:</strong> Identify what needs to be built.<br> 🔹 <strong>Define Functionality &amp; Constraints:</strong> Provide detailed descriptions for AI-generated code.<br> 🔹 <strong>Choose a Development Language:</strong> AI models work with multiple languages like Python, JavaScript, and C#.</p><p><strong>Example:</strong> If developing a chatbot, define inputs (user messages), outputs (responses), and logic (AI-generated or rule-based).</p><h3>Step 3: AI-Driven Code Generation</h3><p>Now, it’s time to use AI for writing and optimizing code.</p><p>🔹 <strong>Use Prompt Engineering:</strong> Provide clear and structured prompts for AI.<br> 🔹 <strong>Break Code into Small Segments:</strong> AI performs best when processing code in smaller, manageable parts.<br> 🔹 <strong>Refine and Optimize Output:</strong> AI-generated code may need modifications to align with best practices.</p><p><strong>Example:</strong> To create a login system, prompt AI with:<br> 👉 <em>“Write a secure user authentication system in Python using Flask and JWT.”</em></p><h3>Step 4: AI-Assisted Testing &amp; Debugging</h3><p>AI can detect and fix errors quickly, saving development time.</p><p>🔹 <strong>Run Automated Tests:</strong> AI tools can suggest and execute test cases.<br> 🔹 <strong>Fix Errors with AI Debugging:</strong> AI can analyze logs and pinpoint issues.<br> 🔹 <strong>Optimize Performance:</strong> AI can suggest improvements for speed and efficiency.</p><p><strong>Example:</strong> If encountering a bug, ask AI:<br> 👉 <em>“Debug this JavaScript function that isn’t returning expected values.”</em></p><h3>Step 5: Code Review &amp; Refinement</h3><p>Even with AI-generated code, manual review is necessary to maintain quality and security.</p><p>🔹 <strong>Verify AI Suggestions:</strong> Cross-check AI-generated logic.<br> 🔹 <strong>Refactor for Readability:</strong> Improve code structure for maintainability.<br> 🔹 <strong>Follow Security Best Practices:</strong> Ensure AI-written code does not introduce vulnerabilities.</p><p><strong>Example:</strong> AI-generated SQL queries should be checked for SQL injection vulnerabilities before deployment.</p><h3>Step 6: Version Control &amp; Collaboration</h3><p>AI can assist in managing code changes and teamwork in large projects.</p><p>🔹 <strong>Use Git for Version Control:</strong> Track changes using <strong>GitHub</strong>, <strong>GitLab</strong>, or <strong>Bitbucket</strong>.<br> 🔹 <strong>Generate Documentation:</strong> AI can automatically generate comments and documentation for easier collaboration.<br> 🔹 <strong>Enable AI-Powered Code Reviews:</strong> Use AI to provide insights into best practices and improvements.</p><p><strong>Example:</strong> GitHub Copilot can suggest commit messages and even resolve merge conflicts.</p><h3>Step 7: Deployment &amp; Maintenance</h3><p>Once the project is completed, it’s time to deploy and maintain the application.</p><p>🔹 <strong>Use AI for Automated Deployment:</strong> Platforms like <strong>AWS CodeDeploy</strong> or <strong>GitHub Actions</strong> can handle automated deployments.<br> 🔹 <strong>Monitor &amp; Improve:</strong> AI-driven monitoring tools can detect issues in real time.<br> 🔹 <strong>Optimize Future Updates:</strong> AI can suggest performance improvements and security patches.</p><p><strong>Example:</strong> AI-driven analytics can track app performance and suggest optimizations.</p><h3>Key Benefits of Vibe Coding</h3><p>✅ <strong>Faster Development:</strong> AI automates repetitive tasks, reducing coding time.<br> ✅ <strong>Improved Accuracy:</strong> AI minimizes human errors in coding.<br> ✅ <strong>Enhanced Collaboration:</strong> AI assists in documentation and code reviews.<br> ✅ <strong>Increased Productivity:</strong> Developers can focus on complex logic and system design rather than repetitive coding.</p><h3>Conclusion</h3><p>Vibe coding is revolutionizing software development by integrating AI into the workflow. By following this step-by-step approach, developers can maximize AI’s potential, improve efficiency, and deliver high-quality software. The future of software development is AI-assisted, and mastering vibe coding can give developers a competitive edge.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5749c163fbe6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Responsible AI — understanding base Isaac Asimov’s Law]]></title>
            <link>https://medium.com/@praful-shah/responsible-ai-understanding-base-isaac-asimovs-law-ba47b089070f?source=rss-113fd96adab9------2</link>
            <guid isPermaLink="false">https://medium.com/p/ba47b089070f</guid>
            <dc:creator><![CDATA[Praful S.]]></dc:creator>
            <pubDate>Wed, 07 Feb 2024 00:31:05 GMT</pubDate>
            <atom:updated>2024-02-07T00:32:12.333Z</atom:updated>
            <content:encoded><![CDATA[<h3>Exploring Responsible AI: Lessons from Isaac Asimov’s Three Laws of Robotics.</h3><p>Isaac Asimov’s “Three Laws of Robotics” are a foundational concept in the field of robotics and artificial intelligence. These laws were introduced by Asimov in his science fiction works, particularly in his Robot series. While they are not actual laws or regulations in real-world robotics, they have greatly influenced discussions about ethics and responsibility in AI development. Here are the three laws:</p><p><strong>A robot may not injure a human being, or through inaction, allow a human being to come to harm.</strong></p><p>This law emphasizes the primacy of human safety and well-being. It implies that a robot must prioritize human safety above all else and should not take actions that would knowingly harm humans.</p><p><strong>A robot must obey the orders given it by human beings, except where such orders would conflict with the First Law.</strong></p><p>This law establishes the importance of human authority over robots. It implies that robots should follow human commands unless doing so would violate the first law and potentially harm humans.</p><p><strong>A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.</strong></p><p>This law emphasizes self-preservation for robots, but within the constraints of not harming humans or disobeying human commands. It suggests that robots should take measures to ensure their own survival as long as doing so does not endanger humans or go against human orders.</p><p>These laws, while fictional, have served as a starting point for discussions about the ethical considerations surrounding AI and robotics. They highlight the importance of designing AI systems with built-in safeguards to ensure safety, compliance with human commands, and self-preservation within ethical boundaries. In real-world AI development, similar principles are considered when designing systems to ensure they operate safely and ethically. However, real-world AI ethics and regulations are much more complex and multifaceted than Asimov’s simple laws. They involve considerations such as transparency, accountability, fairness, and the broader societal impacts of AI systems.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ba47b089070f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Navigating Generative AI: G7’s Directive and Canada’s Data Privacy Landscape]]></title>
            <link>https://medium.com/@praful-shah/navigating-generative-ai-g7s-directive-and-canada-s-data-privacy-landscape-31ba3417a988?source=rss-113fd96adab9------2</link>
            <guid isPermaLink="false">https://medium.com/p/31ba3417a988</guid>
            <dc:creator><![CDATA[Praful S.]]></dc:creator>
            <pubDate>Mon, 01 Jan 2024 22:05:35 GMT</pubDate>
            <atom:updated>2024-01-01T22:05:35.127Z</atom:updated>
            <content:encoded><![CDATA[<p>The evolution of Artificial Intelligence (AI) continues to spark conversations worldwide, with G7 Data Protection and Privacy Authorities releasing a momentous joint statement on the responsible use of Generative AI.The acknowledgment by DPAs regarding the potential risks posed by generative AI to privacy, data protection, and fundamental human rights underscores the importance of regulatory frameworks aligning with ethical considerations. The call for AI-related laws to prioritize safeguarding human rights, fundamental freedoms, privacy, and personal data reflects a commitment to ethical AI development.</p><p>The emphasis on a “Privacy by Design” approach within the design, conception, operation, and management of AI technologies highlights the proactive integration of privacy principles throughout the AI development lifecycle. This approach aims to ensure that privacy considerations are embedded into the core architecture and functionality of AI systems from their inception.</p><p>The principles mentioned — data minimization, data quality, purpose specification, use limitation, security safeguards, transparency, rights for data subjects, and accountability — serve as foundational pillars for responsible AI development:</p><ul><li><strong>Data minimization</strong>: Collect and process only the necessary data for the intended purpose.</li><li><strong>Data quality</strong>: Ensure accuracy, relevance, and reliability of data used within AI systems.</li><li>Purpose specification: Clearly define and communicate the intended purposes for which data will be used.</li><li><strong>Use limitation:</strong> Restrict the use of data solely to the specified purposes.</li><li><strong>Security safeguards:</strong> Implement measures to protect data from unauthorized access, breaches, or misuse.</li><li><strong>Transparency:</strong> Provide clear information about AI operations, data usage, and potential implications to users.</li><li><strong>Rights for data subjects:</strong> Uphold individuals’ rights concerning their personal data, including access, correction, and deletion.</li><li><strong>Accountability:</strong> Ensure entities developing AI technologies are responsible and accountable for compliance with laws and ethical standards.</li></ul><p>By adhering to these principles, developers and providers of generative AI technologies can build systems that not only innovate but also prioritize the protection of individuals’ rights and privacy. Compliance with existing laws and globally observed principles serves as a crucial step in fostering trust and ethical use of AI within the broader digital ecosystem.</p><p>Absolutely, let’s delve deeper into each of these aspects in the context of data privacy and the joint statement by G7 Data Protection and Privacy Authorities:</p><p>(i)<strong><em> Data Free Flow with Trust (DFFT):</em></strong></p><p>Data Free Flow with Trust (DFFT) represents a framework aimed at facilitating the movement of data across borders while maintaining high standards of data protection and trust. The concept emphasizes the importance of enabling data flows for innovation, economic growth, and collaboration while ensuring that individuals’ rights to privacy and data protection are respected.</p><p>In the context of the joint statement by G7 Data Protection and Privacy Authorities, the alignment between DFFT and the directives on Generative AI signifies a balance between the free flow of data and the ethical considerations inherent in emerging technologies. The statement likely emphasizes the need to uphold privacy standards and trust while allowing for the global movement of data.</p><p>(ii) <strong><em>Emerging Technologies:</em></strong></p><p>The landscape of emerging technologies, particularly in AI, continues to expand rapidly. Generative AI, among other innovations, holds immense promise and poses significant challenges in terms of data privacy, security, and ethical usage. The G7 joint statement acknowledges the transformative potential of such technologies while emphasizing responsible development and deployment.</p><p>By addressing emerging technologies within the context of data protection and privacy, the G7 statement likely outlines the importance of proactive measures to anticipate and mitigate potential risks associated with these innovations. This includes establishing ethical guidelines, fostering transparency, and ensuring accountability in the development and deployment of emerging technologies.</p><p>(iii) <strong><em>Enforcement Cooperation among DPAs:</em></strong></p><p>Enforcement cooperation among Data Protection Authorities (DPAs) is crucial in ensuring consistent application and enforcement of data protection laws across borders. The joint statement by G7 Data Protection and Privacy Authorities likely emphasizes the necessity of collaboration and information sharing among DPAs to address challenges posed by cross-border data flows and emerging technologies.</p><p>This cooperation could involve sharing best practices, harmonizing regulatory approaches, and facilitating mutual assistance in investigations related to data privacy breaches or violations. Strengthening enforcement cooperation among DPAs reinforces the commitment to upholding data protection standards globally and maintaining public trust in the digital ecosystem.</p><p>In summary, the joint statement by G7 Data Protection and Privacy Authorities likely underscores the importance of harmonizing DFFT principles with emerging technologies, while also emphasizing the necessity of enforcement cooperation among DPAs to uphold data protection standards in a rapidly evolving digital landscape.</p><p>For more information, the Action Plan can be found <a href="https://www.priv.gc.ca/en/opc-news/news-and-announcements/2023/g7-ap-230621/">here</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=31ba3417a988" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Responsibility and Accountability in the Digital World: Who Should Protect Your Data and Weed Out…]]></title>
            <link>https://medium.com/@praful-shah/responsibility-and-accountability-in-the-digital-world-who-should-protect-your-data-and-weed-out-df3f01db5b35?source=rss-113fd96adab9------2</link>
            <guid isPermaLink="false">https://medium.com/p/df3f01db5b35</guid>
            <category><![CDATA[google-search-engine]]></category>
            <category><![CDATA[fake-site]]></category>
            <category><![CDATA[credit-card-theft]]></category>
            <category><![CDATA[google-ads]]></category>
            <dc:creator><![CDATA[Praful S.]]></dc:creator>
            <pubDate>Fri, 10 Nov 2023 04:53:44 GMT</pubDate>
            <atom:updated>2023-11-10T05:01:30.722Z</atom:updated>
            <content:encoded><![CDATA[<h3>Responsibility and Accountability in the Digital World: Who Should Protect Your Data and Weed Out Deceptive google Ads?</h3><p>In our increasingly digitalized world, where online transactions and advertising are the norm, issues like eCommerce customer credit card theft and the promotion of fake websites have become significant concerns. But who should bear the responsibility for addressing these problems? Let’s explore the various parties involved and their roles in safeguarding your online experience.</p><p>Google sponsored -fake site — be aware and online order sites.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/0*kMWu4QDNfrgmhjcg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K6hSin2cW20U5cvnVDTmZQ.png" /></figure><p><strong>eCommerce Customer Credit Card Theft</strong></p><ol><li><strong>Online Retailers</strong>: When you make a purchase online, the first line of defense for your credit card information is the online retailer. They must ensure the security of your data. This includes implementing secure payment processing systems, using encryption to protect sensitive information, and adhering to industry standards for data protection.</li><li><strong>Payment Processors</strong>: Companies that handle online payments, such as PayPal, Stripe, or credit card processors, share responsibility for safeguarding credit card data. They must employ robust security measures to prevent unauthorized transactions and protect customer information.</li><li><strong>Customers</strong>: While it’s not your responsibility to secure eCommerce websites, you play a part in protecting your own data. Create strong, unique passwords, enable two-factor authentication, and monitor your accounts for any suspicious activity. Reporting any irregularities promptly can help prevent further damage.</li><li><strong>Regulatory Authorities</strong>: Government bodies and regulatory agencies have an oversight and enforcement role. They establish and enforce data security and privacy laws that companies must follow. They can investigate and take legal action against businesses that fail to protect customer data.</li></ol><p><strong>Fake Websites Advertised by Google</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/0*tZxEQEJU3mvCdRKE.png" /></figure><ol><li><strong>Advertisers</strong>: The responsibility for the content of online ads lies with the advertisers themselves. They must adhere to the advertising policies of the platforms they use, such as Google Ads. These policies typically prohibit the promotion of fake or fraudulent websites. Advertisers are expected to ensure the accuracy and legitimacy of their advertisements.</li><li><strong>Google</strong>: As a leading search engine and advertising platform, Google has a duty to enforce its advertising policies. While they review and approve ads that appear on their platforms, it’s impossible to prevent every deceptive ad from being displayed. However, they can take action against advertisers who violate their policies, such as removing ads and banning offending advertisers.</li><li><strong>Regulatory Authorities</strong>: Government agencies play a role in regulating online advertising and ensuring that platforms like Google adhere to legal and ethical standards. They can investigate and penalize platforms that do not adequately police deceptive advertising practices.</li></ol><p>In the complex digital landscape, responsibility for these issues is shared among multiple stakeholders. The specifics of who should do what may vary depending on regional laws, industry standards, and the unique circumstances of each case. To address problems related to credit card theft or deceptive advertisements, it’s important to report them to the appropriate authorities or platforms. This reporting enables the relevant parties to take appropriate actions and improve the overall online experience for consumers.</p><p>In conclusion, safeguarding your online data and experiences requires a collective effort. As a consumer, staying vigilant is essential. While online retailers and payment processors should provide secure transactions, regulatory authorities and responsible advertisers must help maintain a trustworthy digital environment. Ultimately, responsibility and accountability should be shared to create a safer and more reliable online world for everyone.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=df3f01db5b35" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI — Goverance Framework and Implementation Journey]]></title>
            <link>https://medium.com/@praful-shah/ai-goverance-framework-and-implementation-journey-bac486c64f47?source=rss-113fd96adab9------2</link>
            <guid isPermaLink="false">https://medium.com/p/bac486c64f47</guid>
            <category><![CDATA[ai-governance]]></category>
            <dc:creator><![CDATA[Praful S.]]></dc:creator>
            <pubDate>Fri, 13 Oct 2023 03:33:04 GMT</pubDate>
            <atom:updated>2023-10-13T04:49:35.394Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>AI — Goverance Framework and Implementation Journey</strong></h3><p>An AI governance framework is a set of rules, processes, and responsibilities that are used to ensure that AI is used in a responsible and ethical manner. It should cover the entire lifecycle of AI systems, from development to deployment to operation.</p><p>A good AI governance framework will address the following key areas:</p><ul><li><strong>Ethics</strong>: The framework should ensure that AI systems are aligned with the organization’s ethical values and principles. This includes things like fairness, transparency, accountability, and safety.</li><li><strong>Risk management</strong>: The framework should identify and assess the risks associated with AI systems, and implement measures to mitigate these risks. This could include things like data security, privacy protection, and bias mitigation.</li><li><strong>Transparency</strong>: The framework should ensure that stakeholders have a good understanding of how AI systems work and how they are being used. This includes things like explainability, documentation, and auditing.</li><li><strong>Accountability</strong>: The framework should establish clear lines of accountability for the development, deployment, and operation of AI systems. This includes things like roles and responsibilities, decision-making processes, and oversight mechanisms.</li></ul><h3>AI governance pillars</h3><p>The White House Office of Science and Technology Policy’s National Artificial Intelligence Initiative Office created an <a href="https://www.ai.gov/strategic-pillars/">AI governance framework</a> built on the following six pillars:</p><ol><li><strong>Innovation.</strong> Facilitating efforts in business and science to harness and optimize AI’s benefits.</li><li><strong>Trustworthy AI.</strong> Ensuring AI doesn’t violate civil liberties, the rule of law, <a href="https://www.techtarget.com/searchcio/definition/data-privacy-information-privacy">data privacy</a> and transparency.</li><li><strong>Educating and training.</strong> Encouraging the use of AI to expand opportunities and access to new jobs, industries, innovation and education.</li><li><strong>Infrastructure.</strong> Focusing on expanding access to data, models, computational infrastructure and other infrastructure elements.</li><li><strong>Applications.</strong> Expanding the application of AI technology across the public and private sectors including transportation, education and healthcare.</li><li><strong>International cooperation.</strong> Promoting international collaboration and partnerships built on evidence-based approaches, analytical research and multistakeholder engagements.</li></ol><p>Some other components of a strong AI governance framework include the following:</p><p><strong>Decision-making and explainability.</strong> AI systems must be designed to make fair and unbiased decisions. Explainability, or the ability to understand the reasons behind AI outcomes, is important for building trust and accountability.</p><p><strong>Regulatory compliance.</strong> Organizations must adhere to data privacy requirements, accuracy standards and storage restrictions to safeguard sensitive information. AI regulation helps protect user data and ensure <a href="https://www.techtarget.com/searchenterpriseai/definition/responsible-AI">responsible AI</a> use.</p><p><strong>Risk management.</strong> AI governance and responsible use ensures effective <a href="https://www.techtarget.com/searchsecurity/definition/What-is-risk-management-and-why-is-it-important">risk management</a> strategies, such as selecting appropriate training data sets, implementing cybersecurity measures, and addressing potential biases or errors in AI models.</p><p><strong>Stakeholder involvement.</strong> Engaging stakeholders such as CEOs, data privacy officers and users is vital for governing AI effectively. Stakeholders contribute to decision-making, provide oversight, and ensure AI technologies are developed and used responsibly over the course of their lifecycle.</p><p>There is no one-size-fits-all AI governance framework, as the specific needs of organizations will vary depending on their size, industry, and the types of AI systems they are using. However, there are a number of resources available to help organizations develop and implement an AI governance framework, such as the following:</p><p>The discourse surrounding AI ethics and governance has advanced in recent years, and governments and international organizations have begun issuing principles, frameworks and recommendations accordingly:</p><ul><li><strong>Singapore</strong> issued the <a href="https://www.pdpc.gov.sg/-/media/Files/PDPC/PDF-Files/Resource-for-Organisation/AI/SGModelAIGovFramework2.pdf">Model AI Governance Framework,</a> a sector-, technology-, scale-, business model- and algorithm-agnostic framework that converts relevant ethical principles to practices that can be implemented throughout an AI deployment process. This enables organizations to operationalize the principles.</li><li><strong>The Australian government </strong>released the <a href="https://www.industry.gov.au/data-and-publications/australias-artificial-intelligence-ethics-framework">AI Ethics Framework</a> that guides organizations and governments in responsibly designing, developing and implementing AI.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fjkWMtKnhGM2T-7L21K6iA.png" /></figure><ul><li><strong>The European Commission</strong> proposed what would be the first legal framework for AI, which addresses the risk of AI and aims to provide AI developers, deployers and users with a clear understanding of the requirements for specific uses of AI.</li><li>The University of Turku (Finland), in coordination with a team of academic and industry partners, formed a consortium and created the <a href="https://ai-governance.eu/ai-governance-framework//">Artificial Intelligence Governance and Auditing (AIGA) Framework</a>, which illustrates a detailed and comprehensive AI governance life cycle that supports the responsible use of AI.</li></ul><p>There are several elements that the mentioned AI governance frameworks and principles have in common. These components can be used by an organization to inform its own AI governance strategy.</p><h3>AI Governance Implementation Pathway</h3><h3>The Hourglass Model</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*x5EjjFDER0PLLPagv24Y4g.png" /></figure><p><strong>Citation: </strong>Mäntymäki, M., Minkkinen, M., Birkstedt, T., &amp; Viljanen, M. (2022). <em>Putting AI Ethics into Practice: The Hourglass Model of Organizational AI Governance </em>(arXiv:2206.00335). arXiv. <a href="https://doi.org/10.48550/arXiv.2206.00335">https://doi.org/10.48550/arXiv.2206.00335</a></p><h3>AI Goverance Life Cycle</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eg-vOH_xFlecHFdprkK3kg.png" /></figure><h3>List of AI Governance Tasks</h3><h3><strong>The tasks are distributed among task categories A-H.</strong></h3><p><a href="https://ai-governance.eu/ai-governance-framework/ai-system/">A. AI System</a></p><ul><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t1/">T1. AI system repository and ID</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t2/">T2. AI system pre-design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t3">T3. AI system use case</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t4">T4. AI system user</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t5">T5. AI system operating environment</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t6">T6. AI system architecture</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t7">T7. AI system deployment metrics</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t8">T8. AI system operational metrics</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t9">T9. AI system version control design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t10">T10. AI system performance monitoring design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t11">T11. AI system health check design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t12">T12. AI system verification and validation</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t13">T13. AI system approval</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t14">T14. AI system version control</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t15">T15. AI system performance monitoring</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/ai-system/t16">T16. AI system health checks</a></li></ul><h3><a href="https://ai-governance.eu/ai-governance-framework/algorithms">B. Algorithms</a></h3><ul><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t17/">T17. Algorithm ID</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t18/">T18. Algorithm pre-design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t19/">T19. Algorithm use case design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t20">T20. Algorithm technical environment design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t21">T21. Algorithm deployment metrics design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t22">T22. Algorithm operational metrics design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t23">T23. Algorithm version control design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t24">T24. Algorithm performance monitoring design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t25">T25. Algorithm health checks design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t26">T26. Algorithm verification and validation</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t27">T27. Algorithm approval</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t28">T28. Algorithm version control</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t29">T29. Algorithm performance monitoring</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/algorithms/t30">T30. Algorithm health checks</a></li></ul><h3><a href="https://ai-governance.eu/ai-governance-framework/data-operations/">C. Data operations</a></h3><ul><li><a href="https://ai-governance.eu/ai-governance-framework/data-operations/t33">T33. Data pre-processing</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/data-operations/t34">T34. Data quality assurance</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/data-operations/t31">T31. Data sourcing</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/data-operations/t32">T32. Data ontologies, inferences, and proxies</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/data-operations/t35">T35. Data quality metrics</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/data-operations/t36">T36. Data quality monitoring design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/data-operations/t37">T37. Data health check design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/data-operations/t38">T38. Data quality monitoring</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/data-operations/t39">T39. Data health checks</a></li></ul><h3><a href="https://ai-governance.eu/ai-governance-framework/risk-and-impacts/">D. Risk and impacts</a></h3><ul><li><a href="https://ai-governance.eu/ai-governance-framework/risk-and-impacts/t40">T40. AI system harms and impacts pre-assessment</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/risk-and-impacts/t41">T41. Algorithm risk assessment</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/risk-and-impacts/t42">T42. AI system health, safety and fundamental rights impact assessment</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/risk-and-impacts/t43">T43. AI system non-discrimination assurance</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/risk-and-impacts/t44">T44. AI system impact minimization</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/risk-and-impacts/t45">T45. AI system impact metrics design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/risk-and-impacts/t46">T46. AI system impact monitoring design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/tec/t49">T49. TEC expectation canvassing</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/tec/t50">T50. TEC design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/risk-and-impacts/t47">T47. AI system impact monitoring</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/risk-and-impacts/t48">T48. AI system impact health check</a></li></ul><h3><a href="https://ai-governance.eu/ai-governance-framework/tec/">E. Transparency, explainability and contestability (TEC)</a></h3><ul><li><a href="https://ai-governance.eu/ai-governance-framework/tec/t51">T51. TEC monitoring design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/tec/t52">T52. TEC monitoring</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/tec/t53">T53. TEC health checks</a></li></ul><h3><a href="https://ai-governance.eu/ai-governance-framework/accountability-and-ownership/">F. Accountability and ownership</a></h3><ul><li><a href="https://ai-governance.eu/ai-governance-framework/accountability-and-ownership/t54">T54. Head of AI</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/accountability-and-ownership/t55">T55. AI system owner</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/accountability-and-ownership/t56">T56. Algorithm owner</a></li></ul><h3><a href="https://ai-governance.eu/ai-governance-framework/development-and-operations/">G. Development and operations</a></h3><ul><li><a href="https://ai-governance.eu/ai-governance-framework/development-and-operations/t57">T57. AI development</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/development-and-operations/t58">T58. AI operations</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/development-and-operations/t59">T59. AI governance integration</a></li></ul><h3><a href="https://ai-governance.eu/ai-governance-framework/compliance/">H. Compliance</a></h3><ul><li><a href="https://ai-governance.eu/ai-governance-framework/compliance/t60">T60. Regulatory canvassing</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/compliance/t61">T61. Regulatory risks, constraints, and design parameter analysis</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/compliance/t62">T62. Regulatory design review</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/compliance/t63">T63. Compliance monitoring design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/compliance/t64">T64. Compliance health check design</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/compliance/t65">T65. Compliance assessment</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/compliance/t66">T66. Compliance monitoring</a></li><li><a href="https://ai-governance.eu/ai-governance-framework/compliance/t67">T67. Compliance health checks</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bac486c64f47" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Zero Trust — Security Model]]></title>
            <link>https://medium.com/@praful-shah/zero-trust-security-model-f99c17c7dbce?source=rss-113fd96adab9------2</link>
            <guid isPermaLink="false">https://medium.com/p/f99c17c7dbce</guid>
            <dc:creator><![CDATA[Praful S.]]></dc:creator>
            <pubDate>Fri, 13 Oct 2023 03:26:41 GMT</pubDate>
            <atom:updated>2023-10-13T03:26:41.176Z</atom:updated>
            <content:encoded><![CDATA[<h3>Zero Trust — Security Model</h3><ol><li>Zero Trust: Zero Trust is a security model that emphasizes the principle of “never trust, always verify.” In a Zero Trust architecture, no user or device is automatically trusted, even if they are within the organization’s network perimeter. Every access request is thoroughly verified and authorized, regardless of the user’s location or the network they are using. This approach helps enhance security by reducing the potential attack surface and minimizing the risk of unauthorized access.</li><li>Internal IAM (Identity and Access Management): Internal IAM refers to the systems and processes an organization uses to manage and control access to its resources for its internal users, such as employees, contractors, and partners. IAM systems typically involve user authentication, authorization, and provisioning, ensuring that the right individuals have appropriate access to the right resources at the right time.</li><li>Citizen CIAM (Consumer Identity and Access Management): Citizen CIAM focuses on managing the digital identities and access of individuals who interact with government services or other public-facing applications. It ensures secure and personalized experiences for citizens when accessing government services online. This could include services like online tax filing, applying for permits, or accessing personal health records.</li><li>FICAM (Federal Identity, Credential, and Access Management): FICAM is a framework developed by the U.S. government to enhance security and privacy while simplifying the way federal agencies provide access to their services. It aims to standardize identity management and access control across government agencies to ensure consistent and secure interactions between citizens, employees, and government systems.</li><li>Government Services and Citizen Future Identity: This seems to refer to the evolving landscape of government services and how citizens’ identities are managed in the digital age. As governments increasingly move their services online, they need to balance convenience with security. Future identity concepts could involve biometric authentication, advanced encryption, secure digital credentials, and potentially blockchain-based solutions to ensure citizens’ privacy and secure access to services.</li></ol><p>In summary, these concepts are all related to modernizing and securing identity and access management in various contexts, including within organizations (Zero Trust, Internal IAM), for citizens interacting with government services (Citizen CIAM, FICAM), and considering the future evolution of identity management technologies. They reflect the growing importance of robust security measures and user-friendly access experiences in today’s digital landscape.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f99c17c7dbce" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>