LIRNEasia

a) Development of global AI governance mechanisms that enable ENFORCEMENT of accountable AI (i.e. address fundamental challenge of cross border accountability) b) Development of global AI governance processes that are FLEXIBLE enough to keep up with changing technology developments and changing societal concerns/need in the long term c) Development of global AI governance processes though and INCLUSIVE, MULTISTAKEHOLDER processes

• Protection and promotion of human rights
• Safe, secure and trustworthy AI

The given list confuses outcomes, with several necessary (though not sufficient) conditions (or inputs, or intermediary steps/intermediary outcomes) for achieving the end goals/outcomes. And it appears to leave out one key outcome. Outcomes: The selected list are two end goals/outcomes we want from AI governance are a) safe, secure and trustworthy AI; b) and protection of human rights; However, it leaves out a key outcome, that of economic development, which one could argue is of equal importance. Or perhaps the wording could be "enhancing and protection rights and enabling economic development". Or could refer to "human rights inclusive of socio-economic rights" Inputs: Most of the other options you have listed are necessary conditions for achieving the above objectives, not ends in themselves. For example, AI capacity building; interoperability of governance approaches; taking account social/cultural/other aspects/, interoperability of governance mechanisms, and even and transparency, accountability and human oversite will all go a long way towards achieving the good AI outcomes listed above.) are inputs that should help achieve the end objectives/ goals (i.e. they are a means to an end). There are of course other inputs that are required on top of these.

DATA GOVERNANCE: is fundamental to AI governance. And that is not represented. In fact, there's a huge overall between the overnance of data and the governance of AI. MARKET COMPETITION regulation - lock in effects, switching costs, market concentration vs competition, impacts every art of the AI value chain. Market regulation is key part of governance. In fact market regulation is key too that can be used to govern AI and achieve objectives listed and selected above. FINANCING: is a significant challenge, particularly for the majority world. Systems of global digital taxation need to be developed if we are to achieve the financing gap.

Answers built into response in previous question

The AI Dialogue should strive for legally binding governance mechanisms All else is secondary. There are plenty of other mechanisms/forums/talk shops to discuss these matters. What we need is enforceable mechanism to be developed through this AI Dialogue process.

This question has been answered by many other submissions. UN IGF, GPAI, WSIS Reviews, UN IGF, various other civil society processes such as Net Mundial are just some examples. Each has a value and therefore engagement with them can enrich the AI Dialogue and vice versa. The multitude of forums that attempt to discuss AI Governance mean civil society has to be use limited resources to be present everywhere (while private sector can afford the cost of being present everywhere). Therefore mechanisms for liaising with the existing fora/processes and drawing input from them and feeding input back to these existing fora/processes is vita to make the engagement processes seamless for civil society (in particular global majority civil society)

The dialogue should be multistakeholder by design. Instead of two co-chairs from UN member states that now chair the process, three sectoral co-chairs/leads could be appointment: one to represent governments/nation states, one to represent civil society, one to represent private sector. Election/selection of these three co-chairs is up to each sector (i.e. civil society caucus should come together to decide on how to elect their representative; and so on). The co-chairs are responsible for outreach to their sectors; A multi-stakeholder advisory group *roughly* modelled on the UN IGF MAG could (who a government co-chair Each sector could/should be reaching broad agreement on positions, and then come together to 'negotiate' with each other. Despite being inclusive than in the past, many formal UN mechanisms are still dominated by government/nation-state led decision making process. The AI Dialogue convening/event/panels should always be multi-stakeholder by definition. This means "high level panels" and "closed door sessions" with government only would not be allowed.

Citizen consultation like the global surveys done during the SDG design stage could be conducted at regular intervals (e.g. every 5 years?) to gauge the pulse of people, to identify priorities. This can inform the kind of governance we aim for. Online/remote participation could bring in more voices as long as this is done on equal footing to in-person participation. Scientific advisory panel currently set can and should inform the process as well.