AN EVOLUTIONARY MODEL OF AI GOVERNANCE
A Response to the UN Envoy Request for Papers
Our creation of artificial intelligence has unleashed a third evolutionary process that we don't understand and can't control.
Susan Blackmore, The Meme Machine
Respectfully submitted from a KNC collaboration:
• Elizabeth Houlihan, Advisor on Consumer Package Goods for KNC; Managing Partner, The Pathfinder Group of Princeton, LLC
• Betsy McCullar, Advisor on Consumer/Retail for KNC; Partner, The Hilltop Alliance, LLC,
• Kate Newlin, Founder/CEO Kate Newlin Consulting
• Len Tacconi, retired CEO of Health Management Resources, Former CMO for Weight Watchers International
• Sasha Wallinger, Founder Blockchain Style Lab
Introduction
With the advent of this new technology, comes a series of risks which include the non-regulated adoption, lack of understanding about the potential implications that AI exerts throughout the global community, as well as the vulnerability it creates for the UN Member nations across the previously unwitnessed enormity of technological and humanitarian overlaps.
As the Office of the UN Secretary-General's Envoy on Technology prepares for the High-level Advisory Body on Artificial Intelligence (AI), there is momentum to create the prism through which the UN governing bodies focus on key issues, efforts and models to impact AI’s Governance in 2023 and beyond. The paper shares one potential model.
Our proposed model (Fig. A) is a three-dimension pyramid, identifying the risks, challenges and opportunities for collaboration to shepherd the UN’s role in our high stakes future. First, we assess the risks this technology presents. Next, the possibilities it ignites. Then, the challenges it offers the UN as the touchstone of the global community. The best outcome? On-going analysis and the evolution of a strategic roadmap. The goal is to ensure the UN is poised to play foundational and visionary roles in our shared future, thereby benefiting humanity, climate and the many pillars of the Sustainable Development Goals (SDGs) among other strands within the fabric of the future.
Panel One Panel Two Panel Three
Fig. A
1. Panel One: Anticipated Areas of AI Influence in Descending Order of Priority
With the formation of the UN’s High-level Advisory Body on AI comes the opportunity to bring into sharp relief the technology’s societal intersectionality: geographically, socio-politically, and across the cross-section of sectors on which the UN is uniquely empowered to advise. The chart above identifies nine major sectors, organized in order of significance, to outline the many areas AI will impact.
In our view these include the following in descending order of risk:
1. Military, Defense and Weapons
2. Food, Agriculture and Natural Resources
3. Telecommunications and Wireless Technology
4. Government Systems and Data
5. Medical, Healthcare, Pharmaceuticals and Treatments
6. Public and Private Transportation and Vehicles
7. Manufacturing and Industry
8. Consumer goods and services
9. NGOs and Non-Profits & Education
The Advisory Body may have a different view on the order and indeed, the magnitude of risk within these subgroups. Agreement on priorities will, in itself, we suspect, become a critical first step towards aligning on a strategic road map. What is clear is that each of these categories include a series of shared dependencies, including a humanitarian impact, a series of country-specific guidelines and regulations as well as policies and enforcing bodies, including local government, business, regulations and laws.
Within each sector, there are multiple levels of complexity: We believe important early assessment work of the UN’s Advisory Body on AI is to unpack, investigate and gain shared priorities. However, for the purpose of this paper, we rank and outline proposed approaches to model the potential for risk mitigation.
2. Panel Two: Risk Assessment of What Matters Most, First – and for which Level of Governance
The primary priority for the UN’s High-level Advisory Body on AI is to ensure a fundamental understanding of the many critical areas impacted by the escalating global adoption of AI. Each UN Member State and every Observer Nation, using bodies and constituencies will experience radical shifts brought by AI at a different rates and speeds, but each will be touched across virtually every aspect of our daily lives in ways we are only beginning to imagine. Therefore, it is mission critical to identify and prioritize issues and patterns as they emerge across geographies isolating potential working groups to serve as teams to address common challenges and areas of collaboration. This focus results also in increased public and private awareness, vision and oversight as the implementation of AI continues its own evolution, which knows no geographic bounds or limits.
Similar to governing bodies on natural resources, and the formation of the UN SDG’s, we call for an oversight committee to codify and ultimately share a series of unified AI Priorities, driven from this initial sector analysis to act as with which to examine and understand how the future of AI will be met with obstacles that can be bridged by early and evolving guidelines covered by this governing body. We suggest these priorities be formulated by the UN Member States, as a collaboration to ensure the potential for their ongoing adoption and, just like the AI Technology which they will reference, be living, breathing working groups which can nimbly respond to the emerging technology, environmental resources and human-centered needs in real time.
3. Panel 3: Benefit Augmentation
The nine risk sectors we identify also serve as the opportunity areas. There is a syntax to the development of benefits in the model (Fig. B below).
Fig. B
Consider the top priority of military, defense and weaponry. Surely mitigation of risk is the first benefit which AI is poised to deliver. The speed at which AI can identify and track weapon performance anomalies offers tremendous benefit. Next, the ‘creation of positives’ within weaponry systems could include both earlier warning systems and automatic micro-targeting away from high density civilian areas, hospital zones, schools, important heritage sites and critical infrastructure, such as dams, nuclear facilities and others.
Finally, the formation of collaborative relationships sub-phase brings with it the goal of working to forge the AI equivalent of the Geneva Convention, i.e., a shared agreement among member nations – and the military/industrial complex which designs and manufactures for them – to align to sane AI governance principles.
We propose robust discussion of the power of this three-step syntax, implicitly acknowledging as it does the evolutionary nature of an AI roadmap. It cascades throughout each of the nine levels of the model.
Summary:
It is critical from our perspective that the UN undertake this continuing initiative. The UN already has the essential charter, divisions and sub-divisions of organizations in place. Thus, it is uniquely well-poised to apply decades upon decades of global, regional and local knowledge and understanding, augmented as it is by organizational core competencies. We share a model (Fig. C) below identifying potential core competency organizational partnerships.
Fig. C
When endeavoring to draw conclusions and maxims about AI technology there are many unknowns. Each of these offers a glimpse into a troubling, confounding and alternatively thrilling FutureScape. However, it is essential at a minimum to have a shared understanding of the “known unknowns.” Why? Because of the accelerating speed of AI evolution: This is the world’s first technology which is so easily accessible. There is thus clear and compelling need for the UN’s High-level Advisory Body on AI to bring its nation members together in collaboration around these three strategic arenas: Risks, Opportunities, and UN Role. The models to organize these conversations together include:
1. Identify and prioritize the risks AI technology presents to the global community
2. Inventory the sequential development of likely opportunities AI technology can harness across the global divide
3. Agree to the precise charter and near-term plan for the UN to ensure best outcomes for the many.