Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Creating a Cooperative AI Policymaking Platform through Open Source Collaboration

Aiden Lewington, Alekhya Vittalam, Anshumaan Singh, Anuja Uppuluri, Arjun Ashok, Ashrith Mandayam Athmaram, Austin Milt, Benjamin Smith, Charlie Weinberger, Chatanya Sarin, Christoph Bergmeir, Cliff Chang, Daivik Patel, Daniel Li, David Bell, Defu Cao, Donghwa Shin, Edward Kang, Edwin Zhang, Enhui Li, Felix Chen, Gabe Smithline, Haipeng Chen, Henry Gasztowtt, Hoon Shin, Jiayun Zhang, Joshua Gray, Khai Hern Low, Kishan Patel, Lauren Hannah Cooke, Marco Burstein, Maya Kalapatapu, Mitali Mittal, Raymond Chen, Rosie Zhao, Sameen Majid, Samya Potlapalli, Shang Wang, Shrenik Patel, Shuheng Li, Siva Komaragiri, Song Lu, Sorawit Siangjaeo, Sunghoo Jung, Tianyu Zhang, Valery Mao, Vikram Krishnakumar, Vincent Zhu, Wesley Kam, Xingzhe Li, Yumeng Liu

TL;DR

The paper tackles governance challenges posed by rapidly advancing AI, highlighting regulatory lag and unequal societal impacts. It proposes a cooperative policymaking platform built from three pillars—the multimodal Economics Transformer for time-series forecasting, the AI Legislator for value elicitation, and a Policy Interface for transparent interaction—implemented as an open-source initiative. Key innovations include a Continuous-Valued Transformer with diffusion-based learning, a value-elicitation framework grounded in Moral Foundations Theory, and a retrieval-augmented policy generator, complemented by a Dynamic Benchmarking of Indicator Time Series (DBITS) pipeline. Together, these components aim to deliver open baselines, robust forecasts with uncertainty quantification, inclusive policy proposals, and a transparent, trust-building governance tool with broad applicability.

Abstract

Advances in artificial intelligence (AI) present significant risks and opportunities, requiring improved governance to mitigate societal harms and promote equitable benefits. Current incentive structures and regulatory delays may hinder responsible AI development and deployment, particularly in light of the transformative potential of large language models (LLMs). To address these challenges, we propose developing the following three contributions: (1) a large multimodal text and economic-timeseries foundation model that integrates economic and natural language policy data for enhanced forecasting and decision-making, (2) algorithmic mechanisms for eliciting diverse and representative perspectives, enabling the creation of data-driven public policy recommendations, and (3) an AI-driven web platform for supporting transparent, inclusive, and data-driven policymaking.

Creating a Cooperative AI Policymaking Platform through Open Source Collaboration

TL;DR

The paper tackles governance challenges posed by rapidly advancing AI, highlighting regulatory lag and unequal societal impacts. It proposes a cooperative policymaking platform built from three pillars—the multimodal Economics Transformer for time-series forecasting, the AI Legislator for value elicitation, and a Policy Interface for transparent interaction—implemented as an open-source initiative. Key innovations include a Continuous-Valued Transformer with diffusion-based learning, a value-elicitation framework grounded in Moral Foundations Theory, and a retrieval-augmented policy generator, complemented by a Dynamic Benchmarking of Indicator Time Series (DBITS) pipeline. Together, these components aim to deliver open baselines, robust forecasts with uncertainty quantification, inclusive policy proposals, and a transparent, trust-building governance tool with broad applicability.

Abstract

Advances in artificial intelligence (AI) present significant risks and opportunities, requiring improved governance to mitigate societal harms and promote equitable benefits. Current incentive structures and regulatory delays may hinder responsible AI development and deployment, particularly in light of the transformative potential of large language models (LLMs). To address these challenges, we propose developing the following three contributions: (1) a large multimodal text and economic-timeseries foundation model that integrates economic and natural language policy data for enhanced forecasting and decision-making, (2) algorithmic mechanisms for eliciting diverse and representative perspectives, enabling the creation of data-driven public policy recommendations, and (3) an AI-driven web platform for supporting transparent, inclusive, and data-driven policymaking.

Paper Structure

This paper contains 20 sections, 3 equations, 3 figures.

Table of Contents

  1. Introduction
  2. Project Overview
  3. Economics Transformer: Enhancing LLMs for Predicting Policy Impacts on Multivariate Time Series
  4. Data Collection for the Economic Transformer
  5. Novel Architecture Explorations: Foundation Model for Multi-Modal Time Series Tasks
  6. Project Summary
  7. Background and Objectives
  8. Research Direction 1: Fine-Tuning LLMs for Multimodal Ability
  9. Research Direction 2: Designing Time Series Multimodal Model
  10. Scaling Laws for Multimodal Time-Series Language Models
  11. DBITS: Dynamic Benchmarking of Indicator Time Series
  12. AI Legislator: Eliciting Human Preferences for Creation of AI-Generated Policies
  13. Value Elicitation Framework for AI Policymaking: A Simulation-Based Approach
  14. Policy Generation for AI Legislator
  15. Infrastructure for an AI-Driven Policy Interface
  16. ...and 5 more sections

Figures (3)

  • Figure 1: A broad overview of the system’s workflow. The AI Legislator first collects user values and policy objectives from user interactions mediated by the policy interface (arrow 1) and transmits them to both the Policy Interface and the Economics Transformer (arrow 2). The Economics Transformer also receives macroeconomic time-series data from the Federal Reserve, and uses these inputs to generate forecasts, informing the effect of policies on various macroeconomic indicators like GDP over time. These forecasts, along with the proposed policies, are then routed back to the Policy Interface (arrow 3), where policymakers, researchers, and the public can interact with, refine, and better understand policy.
  • Figure 2: Policy Generator Model Frontend Mock-Up.
  • Figure 3: Value Elicitation Model Frontend Mock-Up.