The recent UN Report titled “Governing AI for Humanity” is a comprehensive document that offers a framework for inclusive and beneficial AI governance. Here’s a summary of the main points covered in the report:
Introduction to AI’s Potential and Challenges The report begins by acknowledging AI’s vast potential to benefit humanity, including scientific advancements, resource optimization, and everyday assistance. However, it also highlights the rapid pace of AI development and the concentration of AI capabilities within a few private sector actors and states. From the UN’s perspective, this uneven distribution poses significant risks and challenges, making UN intervention a necessity for a global AI governance approach.
We could easily question the premise of this motivation considering the broadly applicable EU AI Act already weaving its way into the fabric of every decision involving either developing AI tools or applying them to consumer facing problems.
Global Governance Deficit A key concern addressed in the report is the global governance deficit in AI. According to the admittedly biased UN committee, the current landscape lacks a cohesive framework that ensures equal participation, accessibility, representation, oversight, accountability, and prevention of irresponsible AI development. The report emphasizes the UN’s role in addressing this deficit and guiding AI governance.
Those of us in the AI space know that almost all industrialized nations are already presenting AI governing frameworks, so the UN’s perception of a dire need for more regulation may not be accurate.
Guiding Principles for AI Governance: The report outlines five guiding principles for AI governance:
1. Inclusivity: AI should be governed by and for the benefit of all. This matches with the EU AI Act and the executive orders from the Biden administration.
2. Public Interest: AI must be governed in the public interest. This is somewhat of a departure from the EU AI Act because private AI applications are not considered harmful by default, as the UN seems to assert.
3. Data Governance Alignment: AI governance should align with data governance and promote data commons. This is the same approach as all other AI regulations.
4. Universal Collaboration: AI governance must be universal, networked, and rooted in adaptive multi-stakeholder collaboration. This statement is almost too vague to be clearly applicable to most AI developers or those applying AI systems. Universal collaboration is both unrealistic and unnecessary.
5. Anchoring in International Norms: AI governance should be anchored in the UN Charter, International Human Rights Law, and other international commitments like the Sustainable Development Goals (SDGs). This statement asserts the UN’s desire for control over AI, pure and simple.
Institutional Functions for AI Governance The report proposes seven institutional functions to support AI governance:
1. Regular Assessment: Assess the future directions and implications of AI regularly. DSG helps clients globally to assess and control their AI to avoid regulatory infractions as well as performance issues.
2. Interoperability and Norms: Reinforce interoperability of governance efforts and ground them in international norms through a Global AI Governance Framework endorsed by the UN. This is another direct push for the UN to control AI systems through a web of subsidiary regulatory actors. It is likely that both the U.S. and EU, as well as other UN aligned nations, will assign either agencies or regulations to adhere to the UN AI framework.
3. Standards and Risk Management: Develop and harmonize standards, safety, and risk management frameworks. The EU and U.S. approaches include standards including the NIST approach and now the EU approach.
4. Economic and Societal Benefit: Facilitate the development, deployment, and use of AI for economic and societal benefits through international cooperation.
5. International Collaboration: Promote international collaboration on talent development, infrastructure access, dataset building, open-source model sharing, and AI-enabled public goods for the SDGs. This appears to promote Non-Governmental Organizations with whom the UN will support agencies or nations in developing AI resources.
6. Risk Monitoring: Monitor risks, report incidents, and coordinate emergency responses. Potentially, this risk monitoring statement could represent the greatest near-term risk, burden, cost, and oversight requirements. DSG specializes in risk monitoring and risk mitigation to avoid infractions before the regulations impact your business with fines or damage to your reputation.
7. Compliance and Accountability: This section has the unspoken effect of promoting fines and monitoring to ensure compliance and accountability based on established norms. By working with DSG to handle the heavy lifting for avoiding compliance issues, AI companies can focus on core business issues rather than having to watch their back from regulatory attacks.
Conclusion and Next Steps: The report concludes with a call to action for the international community to engage in the proposed governance framework and work towards a future where AI is governed effectively and ethically. The next steps involve further discussions, collaborations, and the development of concrete strategies to implement the recommendations. These next steps suggest combining all multi-national and national frameworks into a UN-controlled master regulatory structure that supplants national sovereignty.
The “Governing AI for Humanity” report is a significant step towards establishing a global governance framework that can harness AI’s potential while mitigating its risks. It appears to be the latest push to merge AI frameworks from the EU and U.S. into one framework the UN can steer, monitor, and control.
This summary captures the essence of the UN Report and its vision for a future where AI is developed and governed with humanity’s best interests at heart, though determining exactly what represents humanities best interests may be relegated to a small cadre of people located in one room in the UN headquarters rather than those actually developing and applying AI.
