Global AI Governance Alliance (GAIGANow)

The launch of a process leading to a Global Agreement on AI Governance - rather than a series of dialogues.

• Safe, secure and trustworthy AI
• Interoperability of governance approaches
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• AI capacity-building

GAIGANow has identified four key issues relating to AI, namely 1. safety, 2. ethics, 3. equity and 4. interoperability. The first and fourth thematic areas match up with the first and fourth GAIGA issues. Of the remaining themes, AI capacity building is closest to equity and Social, economic, ethical, cultural, linguistic and technical implications of AI is closest to ethics.

There is the concept of Red Lines which is a cross-cutting issue and a key element of some of the AI governance frameworks proposed.

Safety: Everyone in the world, and indeed all future generations are threatened by the hugely dangerous and irresponsible lack of global AI safety control. Allowing a headlong race to develop makes a dangerous situation far more dangerous. Safe, secure and trustworthy AI is of paramount importance. Equity: AI capacity building is crucial if the world is to avoid making permanent and indeed exacerbating the current inequity in the world. Ethics: The third theme is problematic in each of its components: employment is starting to be threatened, with young entrants to the job market the first to suffer. If AI is not globally regulated soon, it will have a devastating adverse impact on the job market, threatening the huge role that work has played in providing meaning to people's lives. Economically the technology risks further strengthening the already powerful and creating a them and us economic world. Disinformation is rife and is starting to undermine trust to a serious degree. With the vast majority of AI training being with English material, the culture of the societies with a small number of language speakers is being steadily buried. Interoperability: For those seeking to benefit from AI systems, the lack of a coherent governance and regulation framework is a clear own goal and the sooner such a framework is established the sooner the world will be able to benefit smoothly with the minimum of legal disputes and confusion.

The most useful role the dialogue can play is to establish the dire ned for AI global governance and to launch a process leading to a global AI Treaty. Short of the above, working groups could be usefully established, helping to make real constructive progress before the next Global Dialogue.

Initiatives worth engaging with include: a. The International Dialogue on AI safety (IDAIS) b. The Pro-Human AI Declaration (Future Life Institute) c. The Red Lines Initiative (The Future Society) d. The Articles to be published before the 2026 Global Dialogue on AI Governance. Will be published on Global Governance of AI website (Future of Life Institute and others) global-governance.ai

For each session, states and other stakeholders should be invited to write proposals as to what should be considered in terms of international governance relating to the theme under discussion. For each theme, based upon those submissions, states and key stakeholders should be invited to participate in the debate. These debates should be facilitated as proposed in answer 17. The facilitator could be assisted by proposals / questions from the other participants.

The voices of the Global South have not been heard sufficiently in global discussions of AI governance and are particularly necessary for the discussion of themes such as capacity building and Social/economic/ethical/cultural/linguistic and technical implications of AI. A different voice that has not been heard in AI governance discussions with Governments is that of existential risk. This would be particularly important for the Safe, Secure and Trustworthy AI theme. Also the voice of international business (not just tech companies but multinational companies that use of market AI systems is not heard sufficiently. Essential for the interoperability theme.

A skilled and knowledgable facilitator could orchestrate a relevant dialogue by avoiding set speeches and being more directive, putting points or relevant questions to specific nations / stakeholders. This could be for each of the main thematic issues. This technique was introduced to the UN system some 25 years ago (e.g. with the Commission for Sustainable Development). Full use should be made of the members of the International AI Scientific Panel. They could be seated with their state representatives and asked to speak at relevant moments by the facilitator.

This is a valid request, but the dialogue needs to focus upon international governance rather than national governance and regulation. The latter can best be handled by a best practice website where good practice can be shared - and perhaps evaluated by a suitable panel of experts. Specific policies and practices for international AI governance have been proposed in answer to Q14, namely a. The International Dialogue on AI safety (IDAIS) b. The Pro-Human AI Declaration (Future Life Institute) c. The Red Lines Initiative (The Future Society) d. The Articles to be published before the 2026 Global Dialogue on AI Governance. Will be published on Global Governance of AI website (Future of Life Institute and others) global-governance.ai e. International Governance of Civilian AI: A Jurisdictional Certification Approach (Oxford Martin) f. Architectures of Global AI Governance: From Technological Change to Human Choice.