“AI capabilities have improved radically in recent years,” wrote economists Erik Brynjolfsson, Anton Korinek, and Ajay K. Agrawal in “A Research Agenda for the Economics of Transformative AI,” a recently published working paper. “Our institutions, organizations, skills, and economic models are struggling to keep pace. In this growing gap lie the greatest risks of the coming decade, as well as the greatest opportunities. We urgently need to improve our understanding of AI’s economic implications.”
“As rapid as improvements in AI have been, from mastering Ph.D.-level exams to writing essays and creating art, there are reasons to believe bigger advances will occur within the next few years,” the authors added. “While it is inherently difficult to predict dates for future inventions, we can’t dismiss the possibility that extremely powerful AI systems will be available soon. Such powerful AI systems would transform society. Even current AI technologies have the potential to affect large sectors of the economy.”
In November of 2024, the US National Academies published “Artificial Intelligence and the Future of Work,” a report based on a study — co-chaired by Brynjolfsson — on the current and future impact of AI on the US workforce. “Today, the speed of technological progress is reshaping not just the tools but also the fabric of the workforce and societal structures,” said the report. “The trajectories that Al-enabled futures might take can lead to outcomes of profound benefit or significant disruption.”
AI’s technological progress is often captured by terms like artificial general intelligence (AGI) and artificial superintelligence, that is, AI technologies that will eventually match or surpass human capabilities across just about all cognitive tasks. Transformative AI (TAI), on the other hand, reflects an increasing consensus in policy circles that even if AI doesn’t quite reach human-level cognitive abilities, it will have a very large impact on society potentially comparable to the agricultural and industrial revolutions.
Brynjolfsson, Korinek, and Agrawal define TAI as the kind of AI that increases total-factor productivity growth by at least 3 to 5 times the historical averages. “Such growth may occur because AI facilitates a radical new set of goods, services, or production processes; because AI changes the relative scarcity of inputs, particularly by making cognitive labor significantly more abundant relative to other factors; or because AI creates novel economic organizations and institutions.”
Their study analyzes how TAI will affect economic processes across three interconnected dimensions:
- how TAI affects innovation processes and the creation of new ideas;
- how TAI reshapes the organization of production factors through new business models, market structures, and institutional arrangements; and
- how TAI substitutes for or augments individual factors of production, particularly human labor and cognitive capabilities.
To better understand the key economic challenges TAI represents, the authors identify nine major Grand Challenges that are likely to shape TAI’s trajectory and impact, and for each such challenge they define a few key research questions that should be pursued to better understand their impact. Let me summarize each of these nine Grand Challenges along with their associated research questions.
1. Economic Growth
“Improved technological capabilities are the key drivers of growth in standard economic models.”
- How can economists detect early signs of an AI-driven growth explosion?
- What will be the main bottlenecks for growth?
- How will the role of knowledge and human capital change?
- What new types of business processes and organizational capital will emerge?
2. Invention, Discovery and Innovation
“Since innovation is the primary driver of economic growth, it is important to understand how TAI may transform the nature and extent of innovation.”
- How and where will TAI automate scientific discovery?
- How will the ability to automate experimentation and problem-solving influence the rate of technological progress?
- What are the likely bottlenecks?
- What will be the likely impact on the frequency and quality of innovations and on the rate of economic growth?
3. Income Distribution
“Labor is the main source of income for the majority of the population, and labor markets therefore play a crucial role in income distribution.”
- Will the capabilities of TAI largely displace workers, or will there be areas with rising labor demand?
- How might changing bottlenecks affect the distribution of economic gains across sectors and populations?
- How will this depend on economic policy?
4. Concentration of Decision-making and Power
“The success of ever-larger models suggests the possibility that the AI industry may become increasingly concentrated, while the success of low-cost models of nearly-equivalent performance and the success of open source models could support increased competition.”
- Will TAI be dominated by a single AI system, a small number of comparable systems or a plethora of systems with varying capabilities and strengths?
- How will TAI affect concentration in the rest of the economy?
- Will larger retailers and manufacturers gain an increased competitive advantage over smaller shops and plants?
- Or will AI democratize expertise and lead to flourishing competition?
5. Geoeconomics
“Geoeconomics is an emerging field which examines the use of a country’s economic strength to exert influence on foreign entities to achieve geopolitical or economic goals by leveraging economic instruments like trade policy, investment, and sanctions to advance national interests.”
- How will TAI reshape the economics of deterrence and the balance of power among states?
- How will TAI impact the stability of economic and military alliances and rivalries?
- How will TAI change the economics of cyber warfare and the defense of critical infrastructure?
- Can regulatory frameworks manage the dual-use nature of AI technologies without hindering economic growth?
6. Information, Communication, and Knowledge
“A key determinant of a society’s economic success is how it manages information, communication, and knowledge. Laws, institutions, incentives, and norms that promote the creation and transmission of accurate information tend to boost economic growth.”
- How will TAI affect the quality of information flows?
- How can we create incentives for the production of useful, high-quality information?
- Will information flows provide deeper and broader insights, even novelty or will they be misleading and destructive?
- Could TAI simply overwhelm human-produced content with the sheer quantity of content it produces?
7. AI Safety & Alignment
“AI safety and alignment refer to the challenge of ensuring that AI systems behave consistently with human values and intentions. As AI grows more powerful and autonomous, the economic implications of their safety and alignment become crucial.”
- How do the costs of AI safety and alignment compare to their economic benefits?
- What are the economic incentives for developing safe and aligned AI systems?
- How can we design incentive structures that encourage prioritizing alignment with broader societal goals?
- How can market mechanisms promote the development of safe and aligned AI systems?
8. Meaning and Well-being
In a 1930 essay, English economist John Maynard Keynes wrote about the potential of a future economic problem which he named technological unemployment, that is, “unemployment due to our discovery of means of economising the use of labour outrunning the pace at which we can find new uses for labour.”
“Keynes’ prediction about solving the economic problem raises fundamental questions about human purpose and fulfillment in a TAI world.”
- What can economics contribute to our understanding of meaning and well-being in a world without work?
- How can we analyze the production and distribution of non-monetary sources of fulfillment?
- What is our ultimate objective in a world where machines can perform essentially all work?
- Is it desirable for work to maintain its current societal importance if we achieve Transformative AI?
9. Transition Dynamics
“Optimizing policies and institutions for a world of TAI is not enough. We must also successfully navigate the transition from our current economic institutions, organizations, and processes. As technology advances, bottlenecks are likely to emerge.”
- How does the speed mismatch between TAI and complementary factors affect the rollout of TAI and how can adjustment costs be minimized?
- How can societies prepare for and respond to potential transition crises, e.g., sudden mass unemployment, system failures, or conflicts triggered by TAI developments?
- How can policy interventions — such as targeted retraining subsidies, adaptive regulatory sandboxes, and organizational-innovation grants — minimize the aggregate adjustment costs during the rollout of TAI?
“The transition to an economy shaped by TAI will not follow a predetermined path,” wrote the authors in conclusion. “Some scenarios offer the promise of vastly enhanced wealth, where TAI drives unprecedented productivity, improves social welfare, and distributes benefits fairly. However, without thoughtful management, the outcome could be dystopian, with increased inequality, mass unemployment, social instability, and even catastrophe, leaving many people worse off.”
“This research agenda highlights the key economic questions and encourages researchers to develop the tools necessary to inform policies that maximize positive outcomes. By identifying key economic indicators, anticipating challenges, and advancing this research agenda, we hope to increase the likelihood that TAI will lead to shared prosperity and a sustainable future for humanity.”
