LogaRisk

First of all, whilst I am a technical specialist and former head of risk and regulator, with experience in AI governance and policy (see my linkedin profile https://www.linkedin.com/in/imane-b-48032ba/) I am contacting you as a science fiction writer (and reader), though that cannot be fully dissociated with my deep systemic risk expertise. The outcomes that would make the dialogue a success would be: 1) Investors are bought in and adhere to responsible investing in AI (note: this is different from investing in responsible AI). This means that investors ask guarantees from AI companies about transparency and governance about they do. 2) Insurance companies publishing guidance of what requirements AI models would have to fit with for them to be insured. 3) A clear communication to the general public about the risks from AI and the societal impacts it may have. For that I would strongly recommend organising "reading marathons" of science fictions books such as Azimov series, or "The machine stops" by E.M. Forster or 1984 by Orwell, or Solaris by Lem. A lot of people talk about sci-fi but they have not actually read the books. In Azimov sci-fi, robots are not exactly friendly, but even he imagined at least they would have 3 laws to regulate them. My book (1st in a series) is just 61 pages, but I have received feedback that it is quite informative of what a dystopian AI society would be https://www.amazon.co.uk/Mirror-Show-Error-Last-Intelligence-ebook/dp/B0GLQN33F9 4) A draft charter for willing governments to ratify about regulating AI firms. 5) An organised roadmap to move away from wide-use AI to narrow use-case AI (by that I mean moving from chatbot heavy/consumer heavy AI to useful AI e.g. for healthcare, climate etc)

• Safe, secure and trustworthy AI
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights
• Transparency, accountability, and human oversight

In my opinion, the key urgency is to regulate AI development or at least to move it from non transparent monopolised development to controlled and carefully sequenced rollout. There needs to be an accountability mechanism for AI development. Also, in my opinion AI is a sustainability related risk so needs to be framed under the environmental and social lens, and at the moment it is not, hence not a factor in any sustainability risk management. Also - as a sci-fi writer I would say that we need the dialogue to go beyond the few heavily technical ideologues and include a more variety of profiles that can actually critically think and challenge the status quo about AI.

Yes. As mentioned above, I think AI is a sustainability-related risk. If you are interested, you can check how I framed it here: https://www.linkedin.com/feed/update/urn:li:activity:7450261405720674304/ I think using the momentum for sustainability integration to include AI could constructively help, especially if the lessons learned from the climate risk discussions are taken into account, to avoid some of the pitfalls.

The main impact for my region is 1) the dilution of attention 2) the dilution of real expertise. everyone can *seem* an expert whilst not, which increases systemic risk overall 3) New risks in the financial system 4) Risk of job loss 5) Misinformation Add to that my region is also impacted by the harms AI can pose for children and other vulnerable social categories.

The AI dialogue can help organise and structure i.e. use convening powers to get international agreements on a minimum charter / protocol for safe AI. for example: if UN assemble can agree to ratify a minimum standard for safe, transparent and accountable AI. The AI dialogue can also promote guidelines for investors and asset owners to actually hold the AI firms they invest in accountable

I think the AI dialogue should look to the existing convening for sustainability-related risks and add an AI dimension to these.

The AI dialogue should have chapters in each continent and should prioritise that each chapter feeds to a central team their priorities on AI safety. This could feed to an overall protocol or charter. In addition, a central secretariat could convene different working groups and forums by type of sector, to raise awareness about AI risks and promote a more transparent development. One of these groups should focus on how to get investments into useful collective use AI (e.g. health and climate).

Civil society Sci-fi writers (with relevant background) Philosophers and critical thinker Women Children The global south

Book clubs or movie clubs where an AI related sci-fi book is read or a movie is shown with a live panel and discussion. I think we all need to reclaim our attention and critical thinking, by taking time to reflect and discuss together, rather than being consumed by short snappy videos or content.

You might be interested in this paper: Roberts, H., Taddeo, M., & Floridi, L. "A Framework for Evaluating Global AI Governance Initiatives." Global Policy https://onlinelibrary.wiley.com/doi/10.1111/1758-5899.70164 and this framework https://cyberriskinstitute.org/artificial-intelligence-risk-management/