Mohammed Bin Rashid School of Government

Some suggestions for AI Dialogue will convene in 2026 in Geneva, on 6 and 7 July, and in 2027, in New York (dates TBD), for the future thematic clusters. • Agile Regulations for AI Governance. • Human-centric AI & Innovation Spillovers. Formats: Workshops that provoke thinking rather than panels – focus on deep understanding of what can be done to move forward. Simulations or Scenarios to understand AI regulatory governance challenges • From a global Dialogue point of view – I thank you for this opportunity to hear other voices than technical or legal ones. AI governance needs to be interdisciplinary, embedding systems thinking, as its impacts cross-cut industry, national, and planetary borders. Without ethicists, psychologists, educators, social workers, health experts, economists, etc., we risk silo thinking. This is a design challenge which I hope that the Global Dialogue will manage by ensuring all voices are heard.

Interoperability of governance approaches;Transparency, accountability, and human oversight;AI capacity-building;

You cannot guarantee safe, secure and trustworthy AI hence the discussion is really about how you respond to threats on values. AI governance is about values, and there is no alignment across the world on the same values. For example, national security and research (which are exempt under the EU AI Act) raise concerns, as AI innovations move faster than we can regulate. There are no universal guidelines for research, even though AI is dual-use and industry-crosscutting, and rampant repurposing of AI systems is worrisome. This has implications for the development of open-source software, open data, and open artificial intelligence models, which are happening without foundational global values. We need basic values endorsed across all countries, with people and the planet at the center, starting with early childhood education and embedded in all education and research centers. We often conflate digital literacy with values needed for global governance. I applaud the need for AI literacy but worry that the perception of what this means is too vague. This is a values and literacy challenge. Human Rights also need to be revamped and not force-fitted - for example, we need Posthumous rights, which are not addressed in our existing human rights. The right to a livelihood is not the same as a right to a job, but again, this needs more discussion as AI learns from human skills and data and then replaces them as being more efficient or productive. A human being is not a machine. We need a deep appreciation of what it means to be human. Often, AI governance is implemented as guidelines, and this makes enforceability impossible, especially when government and private sector purposes conflict (for example, a government's mandate is intergenerational public value while the private sector is profits). We do not have enough discussions on trade-offs. It is not the same thing as risk and nor are these simplistic discussions like innovation versus governance. We are using words like transparency and accountability, which are difficult to implement considering third-party providers, AI agents, and multiple humans and agents in the loop. The terms do not account for the complexity of AI systems, making governance difficult to implement. We need a deep understanding of the terminology used in AI governance from a practical perspective, given the rapid development of AI technologies.

• Most of the focus on AI governance is on data, yet the data being used is tacit information, not implicit data, which is the majority of data - this raises questions on the models - hence transparency should be on the provenance of data, training, model, and weights and funding sources. Most of these models are Western-based and, as a result, will ignore the cultural and historical context important to many people in other regions excluded from the development of these technologies. • The definition of AI systems is too restrictive. For example, in the AI Act (Art 3), while it acknowledges the complexities of 'AI systems', it excludes part of the systems like non-adaptive, hard-coded rules and AI-enabled "biological or organic tools. This definition is very narrow and does not consider the future of AI. We need to understand the spirit of AI, replacing human intelligence, and understand the implications of this.

Several ways 1. Loss of skills, critical thinking, knowledge and culture 2. Jobs displacements 3. Social fabric is being stressed 4. Work insecurity will lead to old (or middle)-age poverty 5. Lack of appreciation of what it means to be human 6. Distrust of instututions (governments, businesses others) 7. The entities that control the net are gatekeepers of information - they can decide what to save, distort or manipulate - this is deeply worrying.

We can build a basic understanding of what AI is and isnt, what it can do and cannot do and what it should do and should not do. Without this we cannot have international cooperation on Ai governance.

1. AI simulation ( we are developing a game for policy makers to udnerstand trade-offs and spillovers on AI decision making) 2. Research funding into impacts of AI on society, skills and culture (since AI is a symbol) 3. Discussions on listed companies also working on the defense sector.

Different stakeholders will contribute to systems thinking. This can be roundtables, a design thinking workshop, a future foresight workhsop or a scenario plannign workshop. Hence the dialogue needs more than talking but activities with some tangible outcomes. Happy to help organize the same.

Educators (not researchers), parents, socialogists, psychologist, union workers, for example.

This can be roundtables, a design thinking workshop, a future foresight workhsop or a scenario plannign workshop. Hence the dialogue needs more than talking but activities with some tangible outcomes. Happy to help organize the same.

This can be roundtables, a design thinking workshop, a future foresight workhsop or a scenario plannign workshop. Hence the dialogue needs more than talking but activities with some tangible outcomes. Happy to help organize the same.