Regulatory Constructs and Governance Models for Commercially Deployed AI Applications, and their impact on judiciaries through comparative mixed-methods research

According to the International Telecommunication Union Global Digital Regulatory Outlook 2023 (ITU GDR), regulatory and digital policy benchmarks are severely lacking, with only half of the metrics currently being met by countries globally. Shockingly:

• Only 40% of countries are adequately prepared with legal instruments for digital markets

• A mere 40% have fully functional appeals mechanisms across economic sectors

• Barely 30% require, or effectively utilize, ex-post reviews of sectorial policies

• Just 20% are committed to using public consultations to guide decision-making

• Only 10% require policy rolling reviews

• Only 5% of countries have mature national frameworks equipped for digital markets

The term "need" appears 114 times in the ITU GDR. It is clearly of the utmost importance to research and define metrics and constructs to regulate and govern commercially deployed artificial intelligence (CdAI) applications. What is also needed is an analysis of the impact that technology is having on judiciaries. This research must be conducted through a transformative multidisciplinary team comprised of engineers, lawyers, economists, statisticians, medical professionals, and others, engaging in a simultaneous analysis that transcends national borders. I am interested in joining such a team and ensuring this research is successful.

Globally, it remains evident that experts are ramping up their ongoing national and internal efforts to prepare for the impending wave of challenges posed by the rapid advancement of CdAI. A direct quote from the ITU GDR aptly describes the current state of affairs:

"New regulatory imperatives take centre stage such as inclusiveness, openness and ethics. Technical expertise and academic research from a range of bodies – think tanks, scientific committees –increasingly inform rule-making and decision-making. The old norm of command-and-control with blanket obligations and rules is today ineffective – and actually impedes innovation and investment hampering regulators as they seek to uplift digital markets across the board."

SOURCE: https://www.itu.int/en/ITU-D/Regulatory-Market/Pages/gdro23.aspx

I intend to leverage the exceptional benefits of a comparative mixed-methods convergent design approach to quantify specific attributes, correlations, and metrics within dual-tracks: (1) engineering standards development and (2) legislative initiatives on both a national and international level. Concurrently, I aim to quantitively analyze the impacts on judiciaries of these development tracks. This approach will assess the operational impact and resource requirements on judiciaries by the decisions and actions of engineering standards bodies and legislative and regulatory bodies, while also considering the correlative impacts of outdated laws and controversies that extend beyond the legislative intent of current laws.

This endeavor will include a qualitative analysis of current and former judiciary members and regulators to evaluate the impact on judiciaries from an operational and resource perspective, as well as the impact of inconsistent verdicts and findings from a national, international, or multijurisdictional viewpoint. My attention and focus remain on a dual research and teaching track wherein we are afforded multiple resources and are well-positioned strategically, geographically, and geopolitically for this endeavor.

The realities emphasized within the ITU GDR illustrate the world’s lack of preparedness to regulate CdAI, both internationally and within their own jurisdictions. I perceive many nations as having an opportunity to lead where the U.S., some nations within the EU, and others in Asia may lag significantly behind, due in part to their economies, the operational structure of their governance, and their current election cycles. Given the 2024 elections globally, we have a unique opportunity to gather immense data in support of this research. We can project and witness in real time the impacts of critical events and we can also observe the economic and geopolitical realities influencing these events through the current state of unregulated CdAI. Specific to this opportunity is our ability to model and research the impact that CdAI regulatory and legislative constructs have on judiciaries.

I have also developed outlines of coursework combining engineering and law students in a multidisciplinary setting. I seek to further develop this initiative with your support. While focusing on policy and governance by identifying and analyzing the ongoing actions of national and international standards bodies and legislative initiatives, I aim to create and employ this combined research and course opportunity as a faculty member across multiple disciplines. This combined classroom approach aims to identify correlations and metrics impacting judiciaries from each of the engineering standards and legislative development tracts related to CdAI. Students will identify correlations and disparities across segments such as social media, cybersecurity governance, patents, medical research, logistics, manufacturing, and across academics and curriculums. While in a multidisciplinary setting, they can further identify the impact of CdAI on operations including costs, resource requisites, and the procedural capabilities and limitations on judiciaries.

Through this research, specific identifiable metrics and methodologies will emerge to better assist judicial operations and ensure that legislative measures align with their intended legislative purposes. The correlations and deviations identified through this research will also serve to better diagnose threats and mitigate risks. I welcome your support and input in making this research a success. I believe together we have the expertise and constructs to craft methodologies necessary to successfully implement this research and teaching opportunities and look forward to doing so.

Identified Research Topic:

Regulatory Constructs and Governance Models for Commercially Deployed AI Applications, and their impact on Judiciaries.

Proposed Research Methods:

I propose that we employ a comparative mixed-methods convergent design approach to (1) quantitatively identify correlating procedures of standards-based bodies and legislative bodies on a national and international level impacting judiciaries, and (2) qualitatively identify their impact on judiciaries from a resource and operational basis on a jurisdictional and interjurisdictional level.

Intended Research Question:

What regulatory procedures and enactment methodologies currently deployed in national and international standards bodies (such as ITU/IEEE vs. NIST/ANSI) correlate with national and international legislative bodies (example: State/Federal vs. National/International), and where do their impacts on judiciaries converge or diverge?

Potential Research Sub-questions:

• What constructs would allow for international efforts to codify the sourcing of user generated content (UGC)? What metrics, if embedded into legislation, would reduce judicial resources and decrease the occurrence of cross-border dispute resolution related to sourcing UGC in a licensing or permit-based regulatory construct related to CdAI?

• What disclosure or notice metrics correlate between standards-based rules and legislative initiatives to address monetizing information or incentivizing algorithmic manipulation?

• What feedback loops or procedures exist in both engineering standards bodies and legislative enactments that correlate with a measurable impact on judicial operations? Where do they first converge and last emerge?

• What correlations, if any, do continuing education requirements for engineers/attorneys have on judiciaries? How are correlations between regulations on input data & LLM training results (CdAI) expressed mathematically?

• How have unregulated social media deployments bypassed the need for a broadcast license on a national or international basis? What other correlations exist between the technological standards being utilized and legislative requisites being circumvented?

• How is the deployment/usage of CdAI (both generative/non-generative large language models utilizing either an on-policy and/or off-policy deep reinforcement learning algorithm) impacting the number of parameters within the model in an unregulated and a regulated scenario on a regional, national, or international basis due to legislative actions or inactions? How are the ‘compute’ requisites of a model impacted by regulations?

Expected Research Results and Potential Outcomes:

• I will identify best practice metrics and models applicable in crafting regulatory constructs to deploy oversight across a national and cross-border operational use of CdAI.

• I will identify governance models that are aligned with standards development and legislative enactments to better design communities of practice and build digital membership schemes as requested by and sought after within the UN ITU objectives.

• I will identify methodologies to reduce resource requirements for judicial operations associated with legislative measures not meeting their legislative intent.

• I aim to encourage and support the continued development of on-going and long term sustained research opportunities where this convergent design model allows for explanatory and/or exploratory sequential design models to better address ongoing CdAI as well as educational changes in curriculums.

• I intend to continue to encourage and support opportunist to bring together select groups of engineers/attorneys as well as legislators and judicial members together in round table forums to present and address the on-going nature and analysis of the resulting research.