Most AI governance frameworks still center on the present: Does the AI work? Is the data clean? Is it compliant? All necessary questions, but they overlook the lineage. A fine-tuned or distilled model inherits not only t...
","url":"https://nsmconsultancy.com/f/auditing-your-ais-ancestry-is-no-longer-an-option","title":"Auditing your AI's ancestry is no longer an option","summary":"Most AI governance frameworks still center on the present: Does the AI work? Is the data clean? Is it compliant? All necessary questions, but they overlook the lineage. A fine-tuned or distilled model inherits not only t...","date_modified":"2026-05-12T09:13:25Z"},{"id":"https://nsmconsultancy.com/f/when-ai-models-inherit-behaviour-no-one-put-there","html_content":"Two stories, published days apart, answer that question differently. Together they should change how we think about AI governance.
","url":"https://nsmconsultancy.com/f/when-ai-models-inherit-behaviour-no-one-put-there","title":"When AI models inherit behaviour no one put there","summary":"Two stories, published days apart, answer that question differently. Together they should change how we think about AI governance.","date_modified":"2026-05-05T05:51:56Z"},{"id":"https://nsmconsultancy.com/f/the-return-of-the-syntax-why-natural-language-isnt-enough-for-e","html_content":"The early promise of Generative AI was simple: write what you mean in plain English. No code, no technical training. Then we started engineering the prompts with personas and Chain-of-Thought instructions when we discove...
","url":"https://nsmconsultancy.com/f/the-return-of-the-syntax-why-natural-language-isnt-enough-for-e","title":"The return of the syntax: Natural language isn't enough","summary":"The early promise of Generative AI was simple: write what you mean in plain English. No code, no technical training. Then we started engineering the prompts with personas and Chain-of-Thought instructions when we discove...","date_modified":"2026-04-27T07:42:54Z"},{"id":"https://nsmconsultancy.com/f/from-excel-sheets-to-ai-governance-lessons-from-the-rag-trenches","html_content":"In 2024, I co-developed StyleCheck, an AI application for the newsroom, as part of the Polis LSE JournalismAI Cohort. Our mission: ground an AI in newsroom editorial guidelines to assist journalists in checking for compl...
","url":"https://nsmconsultancy.com/f/from-excel-sheets-to-ai-governance-lessons-from-the-rag-trenches","title":"From Excel Sheets to AI Governance: Lessons from the RAG Trenches","summary":"In 2024, I co-developed StyleCheck, an AI application for the newsroom, as part of the Polis LSE JournalismAI Cohort. Our mission: ground an AI in newsroom editorial guidelines to assist journalists in checking for compl...","date_modified":"2026-04-14T07:59:44Z"},{"id":"https://nsmconsultancy.com/f/emergent-behavior-in-ai-governance-can%E2%80%99t-wait","html_content":"I had started writing an article about emergent behavior in AI agents when Anthropic published the system card for Claude Mythos Preview, the company’s most capable model to date. What they shared illustrates the concern...
","url":"https://nsmconsultancy.com/f/emergent-behavior-in-ai-governance-can%E2%80%99t-wait","title":"Emergent Behavior in AI: Governance can’t wait","summary":"I had started writing an article about emergent behavior in AI agents when Anthropic published the system card for Claude Mythos Preview, the company’s most capable model to date. What they shared illustrates the concern...","date_modified":"2026-04-10T10:50:43Z"},{"id":"https://nsmconsultancy.com/f/mind-your-prompt-words-you-choose-shape-what-ai-tells-you","html_content":"A new study from Gazzetta and ASL19/Factnameh audited six major AI models on Persian-language queries about Iran: ChatGPT, Claude, Gemini, Grok, DeepSeek, Mistral. The same set of 48 prompts was used for all models. The ...
","url":"https://nsmconsultancy.com/f/mind-your-prompt-words-you-choose-shape-what-ai-tells-you","title":"Mind your prompt: words you choose shape what AI tells you","summary":"A new study from Gazzetta and ASL19/Factnameh audited six major AI models on Persian-language queries about Iran: ChatGPT, Claude, Gemini, Grok, DeepSeek, Mistral. The same set of 48 prompts was used for all models. The ...","date_modified":"2026-04-06T06:43:22Z"},{"id":"https://nsmconsultancy.com/f/ai-in-the-newsroom-part-5","html_content":"SEO and AEO
","url":"https://nsmconsultancy.com/f/ai-in-the-newsroom-part-5","title":"AI in the newsroom: Part 5","summary":"SEO and AEO","date_modified":"2026-03-25T16:57:19Z"},{"id":"https://nsmconsultancy.com/f/ai-in-the-newsroom-part-4","html_content":"Liquid content
","url":"https://nsmconsultancy.com/f/ai-in-the-newsroom-part-4","title":"AI in the newsroom: Part 4","summary":"Liquid content","date_modified":"2026-03-25T16:56:18Z"},{"id":"https://nsmconsultancy.com/f/ai-in-the-newsroom-part-3","html_content":"The literacy gap
","url":"https://nsmconsultancy.com/f/ai-in-the-newsroom-part-3","title":"AI in the newsroom: Part 3","summary":"The literacy gap","date_modified":"2026-03-25T16:55:19Z"},{"id":"https://nsmconsultancy.com/f/ai-in-the-newsroom-part-2","html_content":"Are we losing our voice?
","url":"https://nsmconsultancy.com/f/ai-in-the-newsroom-part-2","title":"AI in the newsroom: Part 2","summary":"Are we losing our voice?","date_modified":"2026-03-25T16:54:19Z"},{"id":"https://nsmconsultancy.com/f/ai-in-the-newsroom-part-1","html_content":"Promising but sill limited
","url":"https://nsmconsultancy.com/f/ai-in-the-newsroom-part-1","title":"AI in the newsroom: Part 1","summary":"Promising but sill limited","date_modified":"2026-03-25T16:54:08Z"},{"id":"https://nsmconsultancy.com/f/the-first-us-ai-literacy-framework-vision-without-obligation","html_content":"The US Department of Labor released the first US AI Literacy Framework on February 13, 2026, built around five content areas : understanding AI principles, exploring AI uses, directing AI effectively, evaluating AI outpu...
","url":"https://nsmconsultancy.com/f/the-first-us-ai-literacy-framework-vision-without-obligation","title":"The first US AI Literacy Framework: vision without obligation ","summary":"The US Department of Labor released the first US AI Literacy Framework on February 13, 2026, built around five content areas : understanding AI principles, exploring AI uses, directing AI effectively, evaluating AI outpu...","date_modified":"2026-03-25T16:44:23Z"}]}