From left to right: Olivier Alais (ITU), Wafa Ben Hassine (OHCHR), Boris Inderbitzin (IEC), Cindy Parokkil (ISO), and Daniel Pap (Council of Europe), on the Solutions Stage of the AI for Good Global Summit, 10 July 2026
On 10 July 2026, I joined a panel on the Solutions Stage of the AI for Good Global Summit in Geneva, titled “From Human Rights Principles to ICT and AI Standards: Perspectives from IEC, ISO and ITU.” The session brought together the three organizations behind the 2025 Seoul Statement, alongside the Council of Europe and the Office of the UN High Commissioner for Human Rights (OHCHR), to talk about how international standards bodies are actually putting human rights commitments into effect.
Daniel Pap of the Council of Europe moderated and opened by explaining why his organization was in the chair: its Framework Convention on Artificial Intelligence is the first binding international treaty on AI, and its HUDERIA methodology exists specifically to turn treaty obligations into something usable. He traced the cooperation already running between everyone on stage, ISO and IEC as contributors to the Framework Convention negotiations and to HUDERIA, a Council of Europe and ITU memorandum of understanding dating back to the 1970s, and ongoing work with OHCHR, before turning to the Seoul Statement itself, the joint commitment IEC, ISO, and ITU adopted in December 2025 to weave human rights more deliberately into how international standards get made.
Wafa Ben Hassine of OHCHR opened with the keynote, calling the Seoul Statement a turning point since three of its four commitments touch human rights directly. She walked through where the momentum is actually showing up: the ETSI human rights checklist now in its test phase, the Australian Human Rights Commission’s structured role inside Standards Australia’s own committees, the report from April’s Brussels seminar on human rights and ICT standardisation, and ISO’s capacity building programme pairing national AI regulators with national standards bodies across fifteen countries. Her broader point was that standards are not purely technical instruments, they are normative tools that shape technology upstream, before legislation or courts ever get involved, which is exactly why OHCHR keeps investing in this space, including a new set of draft guidelines on human rights and technical standards proposed just last month.
Standards are a tool that can preserve national sovereignty while still enabling international interoperability.
Boris Inderbitzin from IEC framed the Seoul Statement as serving two audiences at once: externally, a signal of where IEC stands, and internally, a way of telling engineers that human rights are a matter of binding international law, not just ethics or the Sustainable Development Goals. He offered a genuinely useful mental model, five layers stacked on top of each other, binding international law, non-binding soft law, national law, voluntary instruments like international standards, and underneath all of it the raw normative power of whatever technology happens to make possible. Getting those five layers to line up is, in his view, exactly what international standards are for. He illustrated this with AI incident reporting: the threshold that triggers a report and who gets notified is set by national law, but the content of what gets reported is largely the same everywhere, which is why IEC and ISO are working with the OECD on a shared reporting standard.
Cindy Parokkil from ISO framed the Seoul Statement as a package of four commitments that only make sense together, and pushed hard on one distinction: being invited to the table is not the same as participating effectively. That is why ISO pairs its inclusivity work so closely with capacity building. Nine months into its AI Policy and Standards Capacity Building programme, which brings together fifteen developed and developing countries, the results are already concrete: Malaysia has launched a national initiative putting standards at the foundation of its AI governance, Egypt has built ISO/IEC standards directly into its national AI strategy, Uganda references standards in its draft national digital strategy, and the Netherlands cited the programme as a milestone in its own international AI standards strategy. She closed by announcing a new joint ISO/IEC project, launched just two days before the panel, to run a real-world human rights impact assessment of an AI system in a public procurement context, tied to legal frameworks including HUDERIA, and invited the room to get involved.
For ITU’s turn, I walked through the same ground I have been covering all year: standards are not neutral, they shape how technology touches privacy, dignity, access to information, equality, and non-discrimination, which is why translating human rights into technical language matters so much. I shared where our internal survey across ITU-T study groups landed, 84 percent see human rights as relevant to their work, but 40 percent report low familiarity, and integration remains fragmented and mostly ad hoc. I also described what we are building in response: a draft guideline with screening questions to catch risks early, a short human rights considerations paragraph for draft Recommendations, and three training formats, a masterclass, a self paced ITU Academy course, and an instructor led course applying the tools to real ITU-T cases.
What stayed with me from the session as a whole was how consistently everyone landed on the same point despite four very different institutional starting points: human rights considerations only mean something in context, tied to a specific use case and a specific deployment, not bolted on as a generic checklist. Getting engineers, regulators, human rights experts, and civil society into the same room early enough to work that out together is still the hard part, and still the part worth showing up for.