Economics of Transformative AI

Here are the winning Projects from the Sprint:

1. First Prize: The Early Economic Impacts of Transformative AI: A Focus on Temporal Coherence: Investigates how AI's ability to maintain goal-directed behavior over time determines automation potential and economic impacts
2. Second Prize: Evaluating the risk of job displacement by transformative AI automation in developing countries Analyzes job displacement risks using 4 million Claude prompts mapped to occupation data in Brazil's labor market.
3. Third Prize: Economics of AI Data Center Energy Infrastructure: Strategic Blueprint for 2030 Examines economic bottlenecks and policy solutions for AI infrastructure as energy demand triples by 2030
4. Fourth Prize: Economic Feasibility of Universal High Income (UHI) in an Age of Advanced Automation: Evaluates fiscal measures including wealth taxes and AI dividends to fund UHI during mass technological unemployment.

As AI capabilities advance rapidly, we face unprecedented economic questions: How will transformative AI reshape labor markets and productivity? What distributional effects might emerge, and how can we ensure broad-based prosperity? What governance structures and incentive systems might guide beneficial AI development? These questions demand rigorous economic analysis from diverse perspectives.

Organized by Apart Research in collaboration with BlueDot Impact, this sprint builds on BlueDot's "Economics of Transformative AI" course, providing participants with an opportunity to apply economic theory to practical research questions. While the course provides a foundation, the sprint welcomes all economists with strong analytical backgrounds interested in applying their expertise to AI safety challenges.

The sprint follows a three-phase structure: pre-event development of open research questions, an intensive weekend of collaborative research, and post-event opportunities for exceptional projects to receive continued support through our fellowship program. Join us in developing the economic tools and frameworks needed to navigate one of this century's most consequential transitions.

Why This Matters

Critical Economic Questions Demand Answers

The economic implications of transformative AI will reshape markets, institutions, and societies. Without rigorous economic analysis, we risk navigating this transition blindfolded. Traditional economic models may fail to capture the unique dynamics of transformative AI systems, including recursive self-improvement, exponential productivity growth, and unprecedented labor market effects.

Economists Bring Essential Perspectives to AI Safety

The field of AI safety has been led primarily by computer scientists and alignment researchers. Economic expertise is crucial but underrepresented. Economists bring vital methodological tools, including causal inference, game theory, mechanism design, and welfare analysis that can inform both technical alignment work and governance frameworks.

Bridging Research and Policy

Economic insights will directly inform policy responses to AI development. Your research contributions could influence governance frameworks, regulatory approaches, and international coordination efforts around AI. By developing robust economic analyses now, we can shape more effective policies before transformative capabilities emerge.

Challenge Tracks

Distribution Track

How will the benefits and risks of transformative AI be distributed across society? This track examines the distributional implications of advanced AI systems:

• Labor Market Transitions: Analyze potential employment disruptions, wage effects, and skill premium dynamics as AI capabilities expand across domains
• Wealth and Income Distribution: Model how capital concentration, data ownership, and algorithmic deployment might affect inequality
• Geographic Disparities: Examine how AI impacts may differ across regions, countries, and urban/rural divides
• Policy Interventions: Design and evaluate potential policies for ensuring broad-based prosperity, including UBI, skill transition programs, and data dividends

Distribution Track Project Ideas:

• Distribution of AI's Economic Gains: Analyze how the "AI dividend" might be split among stakeholders - companies, workers, consumers, and society.
• Labor Market Simulation: Model which occupations are most vulnerable to AI automation versus which might grow, and simulate the timeline of these transitions.
• UBI in an AI Economy: Evaluate universal basic income as a policy response to AI-driven labor market disruptions, determining economic feasibility and effects.

Growth Track

How might transformative AI reshape economic growth trajectories and productivity? This track explores macroeconomic implications:

• Productivity Modeling: Develop frameworks to predict and measure AI-driven productivity changes across sectors
• Growth Dynamics: Explore scenarios for economic growth under different AI development trajectories
• Innovation Systems: Analyze how AI might transform R&D processes, knowledge creation, and technological diffusion
• Transition Management: Identify potential economic instabilities during rapid technological transition and design stabilizing mechanisms

Growth Track Project Ideas:

• AI-GDP Growth Modeling: Build a model to project GDP growth under different AI progress assumptions, comparing scenarios of slow vs. fast automation to identify conditions for explosive economic growth.
• Automation Bottlenecks & Baumol's Cost Disease: Examine how "automation-resistant" sectors might impact overall economic growth despite AI advances in other areas.
• AI and Innovation: Investigate how AI systems as research tools might accelerate technological progress, potentially breaking through diminishing returns in R&D.

Market Dynamics Project Ideas:

• AI Industry Structure Analysis: Examine whether the AI sector will tend toward oligopoly or remain competitive, based on factors like compute costs and network effects.
• Algorithmic Collusion Simulation: Create a market simulation to test if AI pricing algorithms naturally learn to collude, threatening consumer welfare.
• Data Markets in the AI Economy: Investigate emerging markets for data as a critical AI input, including potential market failures and governance mechanisms.

Three-Phase Structure

1. Pre-Event: Open Problems in AI Economics

Before the sprint weekend, participants gain access to:

• A comprehensive paper on open problems in AI economics developed by our area chairs
• Curated reading materials (5-20 pages) highlighting key concepts and current research
• The Hackbook template to help structure research approaches
• Discussion forums to connect with other participants and begin forming teams

We recommend reviewing these materials before the event to maximize your productive research time during the weekend.

What to Expect

Previous Apart Research sprints have attracted 300-350 participants, and we anticipate even more for this economics-focused event. You'll join interdisciplinary teams of 3-5 members based on shared research interests.

Our hackathons foster a collaborative atmosphere where participants can:

• Develop novel research insights in a compressed timeframe
• Receive feedback from leading experts in economics and AI safety
• Network with researchers, policymakers, and industry professionals
• Produce work that contributes to the growing field of AI economics

To help focus your research efforts, we strongly encourage you to review:

• Open Problems in the Economics of Transformative AI document: This foundational text outlines the key research questions in the field and provides essential context for your sprint projects.
• Research Directions Document: This practical guide translates broad questions into focused, actionable research directions suitable for the sprint format. It's specifically designed to help you develop novel insights within our compressed timeframe.

Your submission must include:

1. A completed project report using the provided template
2. In case, you want to use Overleaf, here is a template for that
3. Link to a public GitHub repository with your code(optional)
4. A brief (2-3 minute) video demonstration (optional but recommended)

Important: Include an appendix called "Prompts Used" in your submission that documents any LLM prompts used in your project. This helps others understand exactly what was done with the LLM and improves reproducibility.

What background knowledge is required? The sprint primarily targets economists with strong analytical backgrounds who can apply their expertise to AI safety questions. While familiarity with AI concepts is helpful, we welcome participants with diverse backgrounds in economics, including macroeconomics, labor economics, public economics, and development economics.

How are teams formed? Teams will self-organize during the first day based on shared research interests. We'll facilitate team formation through structured activities and interest-matching. You're also welcome to form teams in advance with other registered participants.

What are the expected outputs? Teams will produce research briefs (4-8 pages) outlining their approach, methodology, preliminary findings, and directions for future work. The most promising projects may evolve into academic papers or policy briefs following the sprint.

Will compute resources be provided? Yes, we'll provide access to computational resources for teams requiring data analysis, simulations, or modeling tools.

How will projects be evaluated? Projects will be assessed on:

1. Research quality & scalability
2. AI safety relevance & impact potential
3. Technical quality & methodology
4. Innovation & literature foundation

About Apart Sprints

Apart Research organizes research sprints bringing together diverse experts to tackle crucial questions in AI safety. Our sprints have produced numerous published papers, sparked ongoing research collaborations, and helped identify promising directions in AI safety research.

Previous sprints have focused on topics including mechanistic interpretability, AI security, control systems, and multi-agent dynamics. The Economics of Transformative AI Research Sprint represents our commitment to bringing economic expertise into the broader AI safety ecosystem.

Join us in developing the economic frameworks needed to navigate the transition to transformative AI systems.

Friday (April 25, 4:00 PM - 8:00 PM UTC)

• Welcome and introduction to the sprint format
• Keynote presentation on the state of AI economics research: Join our distinguished speaker for insights into the economics of transformative AI, examining current modeling challenges, emerging research directions, and future priorities for the field.
• Team Matchmaking Session: Find your ideal Economics Research Sprint teammates in this structured networking event! Present your skills and research interests, connect with complementary collaborators, and form balanced teams across our Distribution and Growth challenge tracks.
• Social networking opportunity

Saturday (April 26, 9:00 AM - 10:00 PM UTC)

• Focused research and development time
• Expert office hours with mentors throughout the day
  • Office Hours (1:00 - 2:00 PM): Get expert feedback on your research approaches to labor markets, wealth distribution, and policy interventions from mentors specializing in distributional economics.
  • Office Hours (21:00- 22:00): Consult with growth economics experts on your research examining productivity effects, innovation dynamics, and macroeconomic impacts of transformative AI.
• HackTalk: (5:00 - 6:00 PM): Take a break from your project work to gain valuable insights on research methodologies relevant to AI economics, including analytical approaches, modeling techniques, and interdisciplinary frameworks that can immediately enhance your research sprint project.

Sunday (April 27, 9:00 AM - 5:00 PM UTC)

• Continued research development
• General Economics Office Hours (13:00 - 14:00 AM): Final opportunity to discuss your economic models and analytical approaches with our expert mentors before project finalization.
• Project finalization and preparation of submissions

Post-Event

• Submission Deadline (Monday, April 28, 4:00 - 5:00 AM): Final call for Economics of Transformative AI Research Sprint project submissions! Complete your economic analysis, modeling documentation, and presentation materials.
• Award Ceremony (Thursday, May 1, 4:00 - 5:30 PM): The research sprint might be ending, but the impact continues! Don't miss our Lightning Talks featuring winning projects and special guests. Learn which teams will be invited to the fellowship program.