The Institute of Internal Auditors

Firstly, cross-sector, cross-functional participation. Having diverse voices and views represented would provide richer discussions. Secondly, a nuanced view of the various related issues that AI and AI Governance touches (e.g., energy, workforce, technology, security, GRC). Thirdly, anchoring the conversation to ongoing relevant and topical discussions that move the broader policy dialogue forward in some way. Perhaps the discussions may culminate in a written report that covers a summary of the discussion, discoveries made, and a call to action. Fourth, it's important for this conversation to reach stakeholders who normally don't have a seat at the table. While requiring people to have a follow-up event or bring back new insights to their communities may be challenging, we should ask people to further the conversation, so the ideas and discoveries don't stop when the event stops.

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

Harmonizing (/interoperability) is key! When it comes to moving the needle on policy, it's important people are speaking the same language (even if views differ). We also must focus on the transparency, accountability, and oversight piece because those are important pillars to build trust. And, of course, this trust goes hand in hand with safety. Many people are concerned with the direction AI is going, so we have to center safety in the dialogues. I'm also interpreting capacity building as infrastructure development, primarily, which brings in energy and workforce into the discussion while also keeping in mind the skills development piece, which is also important.

Cybersecurity and quantum are important emerging issues (e.g., Anthropic's Mythos), which may merit further exploration.

Differing views on policy approaches and implementation have been divisive and affect our global profession of internal audit. The auditability of these systems is important to support the proliferation and pervasiveness of this important technology; from a safety and security perspective, we need that transparency and focus on integrity to help develop trust around this. Many organizations lack proper internal controls and the governance frameworks, which means projects may fail to deliver value. There's also a question around the ethical underpinning of the AI governance systems that organizations must face. We're a global association, so there are many gaps around the world on maturity, timing, impact, and understanding. Fortunately, The IIA and the rest of the COSO organizations have written guidance titled, "Achieving Effective Internal Control over Generative AI," which addresses risks on data bias, unclear reasoning, and model drift. The guidance emphasizes accountability, integrity, and verification when dealing with AI/AI outputs.

Soft policy thought leadership which puts a stake in the ground and applies its platform to inform concrete and real policy discussions.

COSO, AI Safety Summit, OECD's Integrated Partnership on AI, Partnership on AI (PAI), Current AI, GovAI Coalition. The added value is in the very fact that people are convening and advancing the conversation-- in building and managing the narrative. I'm happy to make introductions as needed to some of the above organizations/initiatives.

Different stakeholders can contribute to the AI dialogue through sharing their perspectives, providing sponsorship, collaborating on thought leadership, pooling resources, and building connections.

Young people, very elderly people, under-resourced people, and differently-abled people to mention a few. One way to include them is by inviting them directly and having a few scholarships for people to join. Another way is to invite people from associations, civil societies, or NGOs so that their views are represented.

Presentations, round table discussions, demonstrations, case studies, interviews, quick learning (we do something at our conferences where people have lightning presentations, which means that they are on stage for no more than 10 minutes and deliver delivering no more than three key points).

COSO's Achieving Effective internal control over generative AI, The Institute of Internal Auditors' international professional practices framework (IPPF), including the global internal audit standards, OECD guidance, ISO, NIST, and more.