{"id":3534,"date":"2026-03-18T06:00:00","date_gmt":"2026-03-18T06:00:00","guid":{"rendered":"https:\/\/nuno.digital\/porque-e-que-a-maioria-das-iniciativas-de-ia-responsavel-falha-e-como-as-corrigir\/"},"modified":"2026-03-19T11:19:22","modified_gmt":"2026-03-19T11:19:22","slug":"porque-e-que-a-maioria-das-iniciativas-de-ia-responsavel-falha-e-como-as-corrigir","status":"publish","type":"post","link":"https:\/\/nuno.digital\/pt-pt\/porque-e-que-a-maioria-das-iniciativas-de-ia-responsavel-falha-e-como-as-corrigir\/","title":{"rendered":"Porque \u00e9 que a maioria das iniciativas de IA respons\u00e1vel falha (e como as corrigir)"},"content":{"rendered":"\t\t
\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Introduction: The Illusion of Responsible AI<\/h2>\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Over the past few years, \u201cResponsible AI\u201d has become a boardroom priority. Organisations publish ethical principles, appoint AI ethics leads, and reference governance frameworks in strategy decks. On paper, it appears that the industry is making real progress.<\/p>

And yet, beneath the surface, most Responsible AI initiatives quietly fail.<\/p>

Not because organisations don\u2019t care.
Not because the technology is immature.<\/p>

But because\u00a0Responsible AI is treated as a layer on top of systems \u2014 rather than a redesign of how those systems work<\/strong>.<\/p>

This is the uncomfortable truth:<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t\t\t\t\t

Responsible AI doesn\u2019t fail at the level of principles \u2014 it fails at the level of execution.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t\t\t\t\t

Drawing on insights from organisational design and AI implementation (including my recent work through the LSE AI Leadership programme), this article explores\u00a0why Responsible AI initiatives fail in practice<\/em>\u00a0\u2014 and more importantly,\u00a0how to fix them<\/em>.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

The Core Misdiagnosis: AI Is Not a Tool \u2014 It\u2019s a Work Redesign Problem\n<\/h2>\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Most organisations approach AI adoption as if they were deploying software:<\/p>

  • Install the model<\/p><\/li>

  • Integrate it into existing workflows<\/p><\/li>

  • Add a \u201chuman-in-the-loop\u201d<\/p><\/li>

  • Publish ethical guidelines<\/p><\/li><\/ul>

    This mindset fundamentally misunderstands the nature of AI.<\/p>

    AI is not a feature. It is a\u00a0reconfiguration of how work gets done<\/strong>.<\/p><\/blockquote>

    When leaders frame AI as a technical upgrade instead of an organisational redesign, predictable friction emerges:<\/p>

    • AI outputs are ignored because they don\u2019t fit real workflows<\/p><\/li>

    • Teams disagree on how to interpret results<\/p><\/li>

    • Human judgement conflicts with machine recommendations<\/p><\/li>

    • Coordination overhead increases rather than decreases<\/p><\/li>

    • Trust and accountability begin to erode<\/p><\/li><\/ul>

      In short:<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t

      \n\t\t\t\t\t\t\t\t\t

      AI fails not because it is wrong \u2014 but because it is misaligned. <\/strong>And Responsible AI fails alongside it.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

      \n\t\t\t\t\t
      \n\t\t\t\t
      \n\t\t\t\t\t

      Why Responsible AI Initiatives Fail<\/h2>\t\t\t\t<\/div>\n\t\t\t\t
      \n\t\t\t\t\t\t\t\t\t

      Let\u2019s unpack the most common failure modes \u2014 and why they happen.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t

      \n\t\t\t\t\t\t\t\t\t

      1. Ethics Teams Are Disconnected from Product and Engineering<\/h3>

      In many organisations, Responsible AI sits in a separate function:<\/p>

      • Legal<\/p><\/li>

      • Risk<\/p><\/li>

      • Compliance<\/p><\/li>

      • Ethics committees<\/p><\/li><\/ul>

        While well-intentioned, this structure creates a critical disconnect.<\/p>

        Ethics teams define principles.
        Product and engineering teams build systems.<\/p>

        But there is often no operational bridge between the two.<\/p>

        What happens:<\/strong><\/p>

        • Ethical guidelines remain abstract and non-actionable<\/p><\/li>

        • Product teams see governance as a blocker, not an enabler<\/p><\/li>

        • Decisions are escalated too late (often post-launch)<\/p><\/li><\/ul>

          Root cause:<\/strong>
          Responsible AI is treated as a\u00a0review function<\/em>, not a\u00a0design input<\/em>.<\/p>

          How to fix it:<\/strong><\/p>

          • Embed governance into product teams (not outside them)<\/p><\/li>

          • Introduce \u201cResponsible AI by design\u201d checkpoints in discovery and delivery<\/p><\/li>

          • Translate principles into\u00a0product requirements and engineering constraints<\/strong><\/p><\/li><\/ul>

            2. No Clear Ownership of AI Behaviour<\/h3>

            One of the most overlooked challenges in AI systems is ownership.<\/p>

            In traditional software:<\/p>

            • Engineers own logic<\/p><\/li>

            • QA owns testing<\/p><\/li>

            • Product owns requirements<\/p><\/li><\/ul>

              In AI systems:<\/p>

              • Who owns the model\u2019s behaviour?<\/p><\/li>

              • Who is accountable for wrong outputs?<\/p><\/li><\/ul>

                Too often, the answer is unclear.<\/p>

                What happens:<\/strong><\/p>

                • Responsibility is diffused across teams<\/p><\/li>

                • Issues are escalated without resolution<\/p><\/li>

                • Trust erodes internally and externally<\/p><\/li><\/ul>

                  Root cause:<\/strong>
                  AI introduces probabilistic behaviour \u2014 but organisations still operate with deterministic ownership models.<\/p>

                  How to fix it:<\/strong><\/p>

                  • Define\u00a0explicit ownership of AI outputs<\/strong>\u00a0(not just systems)<\/p><\/li>

                  • Assign accountability at the product level (e.g. AI Product Owner)<\/p><\/li>

                  • Establish clear escalation paths for model failures<\/p><\/li><\/ul>

                    3. No Metrics for Responsible AI<\/h3>

                    You cannot manage what you do not measure.<\/p>

                    Many organisations claim to prioritise Responsible AI \u2014 but lack the metrics to support it.<\/p>

                    They measure:<\/p>

                    • Accuracy<\/p><\/li>

                    • Latency<\/p><\/li>

                    • Cost<\/p><\/li><\/ul>

                      But not:<\/p>

                      • Fairness<\/p><\/li>

                      • Consistency<\/p><\/li>

                      • Explainability<\/p><\/li>

                      • User trust<\/p><\/li><\/ul>

                        What happens:<\/strong><\/p>

                        • Governance becomes subjective<\/p><\/li>

                        • Trade-offs are made implicitly (often favouring speed over safety)<\/p><\/li>

                        • Responsible AI remains aspirational rather than operational<\/p><\/li><\/ul>

                          Root cause:<\/strong>
                          Responsible AI is not integrated into performance frameworks.<\/p>

                          How to fix it:<\/strong><\/p>

                          • Define\u00a0operational metrics for Responsible AI<\/strong>, such as:<\/p>

                            • Error severity thresholds<\/p><\/li>

                            • Bias indicators across segments<\/p><\/li>

                            • User trust signals (feedback, overrides)<\/p><\/li><\/ul><\/li>

                            • Integrate these into dashboards, OKRs, and release criteria<\/p><\/li><\/ul>

                              4. No Enforcement Mechanisms<\/h3>

                              Policies without enforcement are just documentation.<\/p>

                              Many organisations have:<\/p>

                              • AI principles<\/p><\/li>

                              • Governance frameworks<\/p><\/li>

                              • Risk guidelines<\/p><\/li><\/ul>

                                But no mechanisms to ensure they are followed.<\/p>

                                What happens:<\/strong><\/p>

                                • Teams bypass governance under delivery pressure<\/p><\/li>

                                • Responsible AI becomes optional<\/p><\/li>

                                • Risk accumulates silently<\/p><\/li><\/ul>

                                  Root cause:<\/strong>
                                  Governance is not embedded into delivery systems.<\/p>

                                  How to fix it:<\/strong><\/p>

                                  • Introduce governance into:<\/p>

                                    • CI\/CD pipelines (e.g. eval gates)<\/p><\/li>

                                    • PR reviews (AI behaviour checks)<\/p><\/li>

                                    • Release approvals (risk sign-off)<\/p><\/li><\/ul><\/li>

                                    • Make Responsible AI part of the\u00a0Definition of Done<\/strong><\/p><\/li><\/ul>

                                      5. AI Is Layered Onto Workflows That Were Never Redesigned<\/h3>

                                      This is the most fundamental failure \u2014 and the one most leaders miss.<\/p>

                                      From the LSE perspective:<\/p>

                                      AI failure is rarely about model accuracy \u2014 it is about organisational fit.<\/p><\/blockquote>

                                      Organisations introduce AI into existing workflows without redesigning them.<\/p>

                                      They ask:<\/p>

                                      • \u201cCan AI do this task?\u201d<\/p><\/li><\/ul>

                                        Instead of:<\/p>

                                        • \u201cHow should this workflow change because AI exists?\u201d<\/p><\/li><\/ul>

                                          What happens:<\/strong><\/p>

                                          • AI outputs arrive at the wrong time<\/p><\/li>

                                          • Teams don\u2019t know how to use them<\/p><\/li>

                                          • Additional coordination and rework is required<\/p><\/li>

                                          • Decision-making becomes slower, not faster<\/p><\/li><\/ul>

                                            Root cause:<\/strong>
                                            AI is treated as an add-on rather than a workflow transformation.<\/p>

                                            How to fix it:<\/strong><\/p>

                                            • Redesign workflows around AI capabilities:<\/p>

                                              • When does AI generate outputs?<\/p><\/li>

                                              • Who consumes them?<\/p><\/li>

                                              • How are decisions made differently?<\/p><\/li><\/ul><\/li>

                                              • Map\u00a0task interdependencies<\/strong>, not just individual tasks<\/p><\/li>

                                              • Align AI outputs with real operational rhythms<\/p><\/li><\/ul>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

                                                \n\t\t\t\t\t
                                                \n\t\t\t\t
                                                \n\t\t\t\t\t

                                                The Real Insight: Responsible AI Is an Operating Model Problem<\/h2>\t\t\t\t<\/div>\n\t\t\t\t
                                                \n\t\t\t\t\t\t\t\t\t

                                                If there\u2019s one idea to take away, it\u2019s this:<\/p>

                                                Responsible AI is not a policy problem \u2014 it is an\u00a0operating model problem<\/strong>.<\/p><\/blockquote>

                                                It requires alignment across:<\/p>

                                                • Strategy<\/strong>\u00a0\u2192 What risks matter?<\/p><\/li>

                                                • Product<\/strong>\u00a0\u2192 How are they designed for?<\/p><\/li>

                                                • Engineering<\/strong>\u00a0\u2192 How are they enforced?<\/p><\/li>

                                                • Operations<\/strong>\u00a0\u2192 How are they monitored?<\/p><\/li><\/ul>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t

                                                  \n\t\t\t\t\t\t\t\t\t

                                                  Without this alignment, Responsible AI remains theoretical.<\/p>

                                                  With it, it becomes a\u00a0scalable capability<\/strong>.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

                                                  \n\t\t\t\t\t
                                                  \n\t\t\t\t
                                                  \n\t\t\t\t\t

                                                  A Practical Framework: Making Responsible AI Work<\/h2>\t\t\t\t<\/div>\n\t\t\t\t
                                                  \n\t\t\t\t\t\t\t\t\t

                                                  To move from aspiration to execution, organisations need to operationalise Responsible AI across four layers:<\/p>

                                                  1. Design Layer<\/h3>

                                                  • Translate ethical principles into product requirements<\/p><\/li>

                                                  • Define acceptable vs unacceptable behaviours<\/p><\/li><\/ul>

                                                    2. Build Layer<\/h3>

                                                    • Implement guardrails, constraints, and evaluation pipelines<\/p><\/li>

                                                    • Integrate Responsible AI checks into development workflows<\/p><\/li><\/ul>

                                                      3. Run Layer<\/h3>

                                                      • Monitor AI behaviour in production<\/p><\/li>

                                                      • Track trust, performance, and failure modes<\/p><\/li><\/ul>

                                                        4. Govern Layer<\/h3>

                                                        • Define ownership and accountability<\/p><\/li>

                                                        • Establish escalation and decision-making structures<\/p><\/li><\/ul>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t

                                                          \n\t\t\t\t\t\t\t\t\t

                                                          This is where Responsible AI becomes real \u2014 not in documents, but in systems.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

                                                          \n\t\t\t\t\t
                                                          \n\t\t\t\t
                                                          \n\t\t\t\t\t

                                                          Conclusion: From Ethics to Execution<\/h2>\t\t\t\t<\/div>\n\t\t\t\t
                                                          \n\t\t\t\t\t\t\t\t\t

                                                          Most organisations don\u2019t fail at Responsible AI because they lack intent.<\/p>

                                                          They fail because they treat it as:<\/p>

                                                          • A compliance requirement<\/p><\/li>

                                                          • A documentation exercise<\/p><\/li>

                                                          • A late-stage review<\/p><\/li><\/ul>

                                                            Instead of what it truly is:<\/p>

                                                            A fundamental redesign of how decisions are made in AI-powered systems.<\/p><\/blockquote>

                                                            The organisations that succeed will be those that:<\/p>

                                                            • Embed governance into product and engineering<\/p><\/li>

                                                            • Redesign workflows around AI<\/p><\/li>

                                                            • Measure what matters<\/p><\/li>

                                                            • Assign clear ownership<\/p><\/li><\/ul>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t

                                                              \n\t\t\t\t\t\t\t\t\t

                                                              In doing so, they won\u2019t just reduce risk. They\u2019ll build something far more valuable:<\/p>

                                                              Trust at scale.<\/strong><\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t

                                                              \n\t\t\t\t\t
                                                              \n\t\t\t\t
                                                              \n\t\t\t\t\t

                                                              FAQs<\/h2>\t\t\t\t<\/div>\n\t\t\t\t
                                                              \n\t\t\t\t\t\t\t
                                                              \n\t\t\t\t\t\t
                                                              \n\t\t\t\t\n\t\t\t\t\t

                                                              1. What is Responsible AI in simple terms? <\/h3><\/span>\n\t\t\t\t\t\t\t\n\t\t\t<\/path><\/svg><\/span>\n\t\t\t<\/path><\/svg><\/span>\n\t\t<\/span>\n\n\t\t\t\t\t\t<\/summary>\n\t\t\t\t
                                                              \n\t\t
                                                              \n\t\t\t\t\t
                                                              \n\t\t\t\t
                                                              \n\t\t\t\t\t\t\t\t\t

                                                              Responsible AI refers to the design, development, and deployment of AI systems in a way that is ethical, transparent, fair, and accountable. It ensures that AI systems align with human values and organisational goals while minimising risk.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/details>\n\t\t\t\t\t\t

                                                              \n\t\t\t\t\n\t\t\t\t\t

                                                              2. Why do most Responsible AI initiatives fail? <\/h3><\/span>\n\t\t\t\t\t\t\t\n\t\t\t<\/path><\/svg><\/span>\n\t\t\t<\/path><\/svg><\/span>\n\t\t<\/span>\n\n\t\t\t\t\t\t<\/summary>\n\t\t\t\t
                                                              \n\t\t
                                                              \n\t\t\t\t\t
                                                              \n\t\t\t\t
                                                              \n\t\t\t\t\t\t\t\t\t

                                                              Most initiatives fail because organisations treat Responsible AI as a policy or compliance exercise rather than embedding it into product design, engineering workflows, and operational processes.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/details>\n\t\t\t\t\t\t

                                                              \n\t\t\t\t\n\t\t\t\t\t

                                                              3. Is Responsible AI only relevant for large organisations? <\/h3><\/span>\n\t\t\t\t\t\t\t\n\t\t\t<\/path><\/svg><\/span>\n\t\t\t<\/path><\/svg><\/span>\n\t\t<\/span>\n\n\t\t\t\t\t\t<\/summary>\n\t\t\t\t
                                                              \n\t\t
                                                              \n\t\t\t\t\t
                                                              \n\t\t\t\t
                                                              \n\t\t\t\t\t\t\t\t\t

                                                              No. While regulation often targets large organisations, any company using AI systems must consider risks such as bias, incorrect outputs, and lack of transparency. Responsible AI is relevant at any scale.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/details>\n\t\t\t\t\t\t

                                                              \n\t\t\t\t\n\t\t\t\t\t

                                                              4. How can product managers contribute to Responsible AI? <\/h3><\/span>\n\t\t\t\t\t\t\t\n\t\t\t<\/path><\/svg><\/span>\n\t\t\t<\/path><\/svg><\/span>\n\t\t<\/span>\n\n\t\t\t\t\t\t<\/summary>\n\t\t\t\t
                                                              \n\t\t
                                                              \n\t\t\t\t\t
                                                              \n\t\t\t\t
                                                              \n\t\t\t\t\t\t\t\t\t

                                                              Product managers play a key role by:<\/p>

                                                              • Defining acceptable AI behaviours<\/p><\/li>

                                                              • Prioritising trust and user experience<\/p><\/li>

                                                              • Embedding governance into product requirements<\/p><\/li>

                                                              • Ensuring accountability for AI outputs<\/p><\/li><\/ul>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/details>\n\t\t\t\t\t\t

                                                                \n\t\t\t\t\n\t\t\t\t\t

                                                                5. What are examples of Responsible AI metrics? <\/h3><\/span>\n\t\t\t\t\t\t\t\n\t\t\t<\/path><\/svg><\/span>\n\t\t\t<\/path><\/svg><\/span>\n\t\t<\/span>\n\n\t\t\t\t\t\t<\/summary>\n\t\t\t\t
                                                                \n\t\t
                                                                \n\t\t\t\t\t
                                                                \n\t\t\t\t
                                                                \n\t\t\t\t\t\t\t\t\t

                                                                Examples include:<\/p>

                                                                • Bias detection across user segments<\/p><\/li>

                                                                • Error severity and frequency<\/p><\/li>

                                                                • User trust indicators (feedback, overrides)<\/p><\/li>

                                                                • Consistency of outputs over time<\/p><\/li><\/ul>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/details>\n\t\t\t\t\t\t

                                                                  \n\t\t\t\t\n\t\t\t\t\t

                                                                  6. What is the difference between AI governance and Responsible AI? <\/h3><\/span>\n\t\t\t\t\t\t\t\n\t\t\t<\/path><\/svg><\/span>\n\t\t\t<\/path><\/svg><\/span>\n\t\t<\/span>\n\n\t\t\t\t\t\t<\/summary>\n\t\t\t\t
                                                                  \n\t\t
                                                                  \n\t\t\t\t\t
                                                                  \n\t\t\t\t
                                                                  \n\t\t\t\t\t\t\t\t\t

                                                                  AI governance refers to the structures, processes, and policies used to manage AI systems. Responsible AI is the broader goal of ensuring those systems behave ethically and safely. Governance is the mechanism; responsibility is the outcome.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/details>\n\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"

                                                                  A maioria das iniciativas de IA Respons\u00e1vel falha \u2014 n\u00e3o por causa da tecnologia, mas por desalinhamento organizacional. Aprenda os principais modos de falha e como operacionalizar eficazmente a IA Respons\u00e1vel.<\/p>","protected":false},"author":1,"featured_media":3535,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"googlesitekit_rrm_CAow74HBDA:productID":"","footnotes":""},"categories":[338],"tags":[364,488,344,1069,331,340,1067,493,1068,378],"class_list":["post-3534","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ia-responsavel-governacao","tag-estrategia-de-produto-de-ia","tag-etica-em-ia","tag-gestao-de-risco-da-ia","tag-governacao-da-aprendizagem-automatica","tag-governacao-da-ia","tag-ia-responsavel","tag-implementacao-de-ia","tag-lideranca-em-ia","tag-modelo-operacional-de-ia","tag-transformacao-de-ia"],"_links":{"self":[{"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/posts\/3534","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/comments?post=3534"}],"version-history":[{"count":2,"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/posts\/3534\/revisions"}],"predecessor-version":[{"id":3538,"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/posts\/3534\/revisions\/3538"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/media\/3535"}],"wp:attachment":[{"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/media?parent=3534"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/categories?post=3534"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/nuno.digital\/pt-pt\/wp-json\/wp\/v2\/tags?post=3534"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}