Subject: What I'm Tracking at the IETF This Week — and Why It Matters to Every Internet User in North America Dear NARALO Members, This week I am attending the IETF's AI Preferences (AIPREF) Working Group interim meetings — a three-day session running Tuesday through Thursday — because I believe our community deserves someone actively monitoring the technical standards that will shape how AI interacts with our content, our data, and our rights. I want to share what I am learning and why I think NARALO needs to pay close attention. The Problem Being Solved Right now, internet users face an "All or Nothing" problem. AI companies constantly scan the web to train their models and power AI search engines. If you want to protect your content from being scraped for AI training, your only real option today is to block bots entirely — which also hides your website from regular search engines. You either give everything away or disappear from the internet. The AIPREF Working Group is building a standardized vocabulary — a set of digital tags — that would let website owners and content creators say, "You can index my content for search, but you cannot use it to train your AI model." That kind of granular control does not exist today. What I Am Seeing in the Sessions The debates happening this week go straight to the heart of end-user protection: The "search" category is facing significant pushback. Four separate issues on the agenda challenge whether the current definition is too broad, whether it redefines established search behavior, and whether it could produce unintended consequences for users. A proposal on the table would narrow the focus to "substantial" generative use — which is exactly the line content creators need drawn clearly. There is a critical shift underway from when content is scraped to what it is used for. This matters enormously. A purpose-based framework gives users meaningful control over their work, rather than a timestamp that tells them nothing about how their content ended up inside an AI model. New categories for Retrieval-Augmented Generation (RAG) and "substitutive use" are being debated. Substitutive use is when AI-generated output effectively replaces a visit to the original source — think of an AI summary that means no one clicks through to the article. How this category gets defined will determine whether creators retain any value from their own work. The Accessibility Risk While reviewing the agenda and tracking the discussions, I identified a concern I believe our community must raise. If we create opt-out tags for AI, we must ensure they do not accidentally break accessibility tools. Many screen readers, on-device translation apps, and assistive summarization tools rely on similar retrieval and generation patterns as AI systems. If an opt-out tag blocks these tools, we create digital inclusion barriers for the people who need the internet most. The RAG and grounding category being debated this week is exactly where this risk is concentrated. I intend to track how this line gets drawn and ensure the accessibility implications stay visible to our community. As someone who recently co-led the VSIG Bengali localization project — where we experienced firsthand how Email Address Internationalization failures blocked participation in our University of Dhaka pilot — I can tell you that these technical gaps are not abstract. When standards miss the inclusion dimension, real people get shut out. Why I Am Telling You This I am running for the NARALO ALAC position because I believe our community needs a representative who does not wait for technical standards to become policy problems. The debates happening at the IETF this week will shape AI governance for years. By the time these decisions surface at ICANN, the technical architecture will already be locked in. My approach is to be in the room — or on the call — where these decisions are being made, translate them into language our community can act on, and bring the end-user perspective into spaces where it is often missing. This is not new for me. Through my work on the RSSAC Caucus, my contributions to the VSIG localization effort, and my ongoing engagement with IETF and ICANN processes, I have consistently worked to connect technical standards to the people they affect. This email is an example of what that representation looks like in practice. I will continue monitoring these sessions through Thursday and will share outcomes with the community. If any of you have direct experience with accessibility tools being blocked by bot-protection measures, I would especially value hearing about it — those real-world examples are exactly what standards bodies need to hear. Best regards, Mohibul Mahmud NARALO Member | ALAC Candidate