University of Oxford

In my opinion, a successful dialogue on AI Governance depends on the substance and post-event implications. There are thousands of discussion forums, but ended up with a non-executable paperworks or no follow up actions. If this Global Dialogue on AI Governance is able to sustain the discussion, produce concrete action points, monitor and evaluate the implementation - I believe it will undoubtedly be a success.

• Safe, secure and trustworthy AI
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Interoperability of governance approaches
• Transparency, accountability, and human oversight

One of the biggest challenges on AI is ensuring it will be aligned with human values, ethically. This is where an AI Governance is critically important. The potential misuse of AI from basic to advanced cases will be able to disrupt many aspects, including to society, economy, and politics. Hence, prioritize the trustworthiness, multidimensional implications, human values, and governance approach are very fundamental to strengthen the foundation of AI implementation and scale the further impacts.

Yes. The listed themes have covered most of topics on AI, but it will be better to capture risk management aspect, industry specific as case study, labor market, policy and stakeholders in building AI Governance. This would help us to ensure the outcomes of discussion is well-balanced between theory and practice.

In my country, Indonesia, the gap is relatively high. For example in financial industry as it is a highly regulated sector, but the evolution of AI technology has challenged the industry to become more adaptable. However, the adoption process or learning curve is quite slow as we are dealing with legacy system, complex bureaucracy, and traditional mindsets. Despite these challenges, a rapid development of AI was encouraged us to accelerate the process, in particular for potential job replacement. It creates both opportunity for the organization to operate better but challenge the people to enhance their skills.

I believe by involving global stakeholders, the AI Dialogue would be able to align the vision and formulate common goals as the facilitator. The discussion is crucial, but again, most importantly is the substance itself.

There are some fellowships program on AI, for example, ERA Cambridge that focusing on AI Safety, AI Technical Governance, and AI Governance. The AI Dialogue can collaborate with a like-minded community to strengthen thought-leadership and work on the same pathways in responding to AI.

It should be a diverse stakeholders from relevant background such as government, industry players, students or experts. A curated topic should be clear along with the objective and expected outcomes. The dialogue can also combine a large conference, fireside chat, role play, and followed by series community of practices.

In my opinion, AI governance is a quite niche topic and not everyone with no expertise would be able to contribute or add values into the discussion. However, if we can bring more industry actors (esp. from conventional organization), children at certain age level who impacted by AI, and students - it can create a dynamic conversation both ideally speaking and reality-check on the ground level.

If there is an interactive technology such as simulator to enhance user or audience experience in AI. This is not only a one-way presentation, but let the participant experiencing the impact of AI. We can also build some scenarios of role play to enable participants share their ideas from diverse background.

Depends on the use case. I believe that AI Governance and policy design would be quite complex, especially at technical level. For example, to enable a chat-bot at banking website or digital platforms, there are some limits due to regulatory barriers that we cannot provide product recommendations to our customer. Hence, we can generate insights, but let customer decide at the end.