The Accelelerators newsletter
Join the Brookings Institution’s AI and Emerging Technology Initiative and Stanford’s Institute for Human Centered AI (HAI) for a conversation on the 2026 edition of HAI’s AI Index and what it reveals about the state of AI development, adoption, and impact.
Closing the data gap for AI policy: Lessons from the Stanford AI Index
OECD POLICY BRIEF
Bridging the AI skills gap: Is training keeping up?
Posted April 24, 2025 Reposted April , 2026
According to the Organisation for Economic Co-operation and Development (OECD), Artificial intelligence (AI) is becoming increasingly important in the workplace. The spread of AIis driving a heightened need for both specialised AI professionals and workers with a moregeneral understanding of AI.
An increasing number of OECD countries are implementing strategies and policies to supportupskilling and reskilling for AI adoption, including the introduction of publicly-funded AI‐related training programmes. However, more can be done to ensure training programmes arewell suited to prepare adults for the future AI-driven workplace.
Analysis from a subset of OECD countries reveals that only a small percentage of trainingcourses currently deliver AI content, signalling a need to scale-up supply.
The majority of programmes with some AI content currently focus on advanced AI skills. Mostcountries could benefit from offering a broader range of courses to promote general AIliteracy.
Policymakers should consider expanding and better targeting initiatives to develop bothgeneral AI literacy and advanced AI skills, including using financial and non-financial incentives,collaboration with industry, and the development of more inclusive learning pathways
Upskilling for the Twin Transitions: Training Supply for Green and AI Shift
The report, Training Supply for the Green and AI Transitions, examines how adult learning systems in OECD countries are adapting to the labor market shifts caused by the "twin transitions" of environmental sustainability and artificial intelligence. While much research focuses on what skills employers demand, this report highlights the supply—whether training providers are actually offering the necessary courses and if those courses are accessible.
## Key Takeaways
The Supply-Demand Gap: * Green Skills: While 20% of workers are in green-driven occupations, green-related training only accounts for 2.1% to 14.1% of available courses.
AI Skills: Despite one in three job vacancies having high exposure to AI, training supply is even lower, ranging from only 0.3% to 5.5% of available courses.
Delivery Methods: * Green training is predominantly in-person due to the need for "hands-on" learning with physical technologies and renewable energy equipment.
AI training leans heavily toward online delivery and is currently skewed toward advanced technical skills for professionals rather than general AI literacy for the broader workforce.
Government Incentives: Most OECD countries have established strategies for the green transition, often providing subsidies or tax deductions for green training. However, financial incentives for AI training are less common and often lack a direct focus on specific AI skill development.
Inclusivity and Vulnerability: Low-skilled adults and workers in carbon-intensive industries are at the highest risk of displacement. While 21 countries offer financial incentives for green training to these groups, only 5 countries have specific career guidance initiatives tailored to the AI transition.
## Comparison of Training Supply
FeatureGreen Transition TrainingAI Transition TrainingCourse Availability2.1% – 14.1%0.3% – 5.5%Primary FormatIn-person / Hands-onOnline / VirtualCurrent FocusVocational & Technical sectorsAdvanced AI ProfessionalsPolicy MaturityHigh (Most countries have strategies)Emerging (Strategies often lack specificity)
### Conclusion
The OECD concludes that the current supply of training is insufficient to meet the rapid pace of the green and AI transitions. To prevent widespread labor market mismatches and rising inequality, policy makers must:
Scale up the volume of AI and green-related training.
Broaden AI training beyond technical experts to include "AI literacy" for the general public.
Improve accessibility for vulnerable workers through targeted career guidance and financial support, ensuring that those most at risk of job displacement are not left behind.
Brookins Institute AI * Technology News
Commentary
It was never the keyboard
Why domain knowledge, not digital skills, determines AI productivity
Posted April 3, 2026 Reposted April 12, 2026
As AI interfaces converge on natural language, the key barrier to AI productivity is domain knowledge.
Domain expertise is appreciating, because it is the input that AI cannot supply: The ability to direct and evaluate its output.
LLMs democratize text production. They do not democratize judgment.
Sacha Alanocaand Chinasa T. Okolo
Posted April 2, 2026 Reposted April 12, 2026
India’s AI Impact Summit signaled a pivot toward a less Western-centric agenda by championing “middle powers” and AI sovereignty as a third path of influence against the traditional global order.
Despite recording 600,000 participants, the summit faced criticism for corporate capture and the physical exclusion of civil society.
As the official AI summit moves to Geneva next year, advocates are calling for a course correction to ensure that the agenda prioritizes genuine stakeholder solidarity over the interests of private corporations.
Economic Round-Up:
AI, Policy, and Market Dynamics
The Future of AI and the Labor Market
Moderate Growth Projections: Economists predict that while AI's effects will be significant, they may not be unprecedented. A "rapid AI" scenario could raise real GDP growth to 3.5% by 2050, though demographic and geopolitical constraints may offset some gains.
Labor Force Shifts: AI is projected to cause job losses particularly among blue-collar workers, potentially contributing to a drop in the labor force participation rate to 55% by 2050.
Adaptive Capacity: Research indicates that while 70% of workers in AI-exposed roles have the capacity to transition, roughly 6.1 million workers—86% of whom are women—lack this adaptive capacity.
Monetary Policy and Financial Stability
Costs of Bank Regulation: Tighter-than-expected bank regulation has been found to reduce lending and raise unemployment for several years, suggesting higher short-run macroeconomic costs than previously estimated.
Housing Market Stabilization: Contractionary monetary policy is found to be at least twice as effective at mitigating housing market overvaluation as expansionary policy is at exacerbating it.
Global Shocks: Global shocks now play a much larger role in driving interest rate movements than in previous decades, with their contribution doubling since 1999.
Trade and Global Competition
Tariff Implications: Recent Supreme Court rulings striking down certain tariff authorities have left businesses to reassess trade paths. Experts note that tariffs on intermediate goods (inputs) generally have more severe economic effects than those on final consumer goods.
The "China Shock": Research suggests China's currency peg amplified the "China shock" between 2000 and 2012; a floating exchange rate might have eliminated 59% of the resulting U.S. manufacturing decline.
USMCA Review: The U.S.-Mexico-Canada Agreement is undergoing its first joint review to determine whether to renew, revise, or terminate the pact.
Productivity and Labor Supply
Manufacturing Quality: Official measures may be understating manufacturing productivity growth by failing to sufficiently adjust for quality improvements, particularly in high-tech sectors.
Medicaid and Work Hours: The decline in average hours worked in the U.S. relative to other advanced economies is partially attributed to an increase in publicly provided health benefits like Medicaid for non-employed individuals.
Apprenticeship Barriers: High administrative burdens and inconsistent standards continue to deter U.S. employers from launching needed apprenticeship programs.
Next Step: Would you like me to draft a specific introductory paragraph for this round-up or perhaps format it as a ready-to-send Gmail draft?
The Future of AI and the Labor Market
Moderate Growth Projections: Economists predict that while AI's effects will be significant, they may not be unprecedented. A "rapid AI" scenario could raise real GDP growth to 3.5% by 2050, though demographic and geopolitical constraints may offset some gains.
Labor Force Shifts: AI is projected to cause job losses particularly among blue-collar workers, potentially contributing to a drop in the labor force participation rate to 55% by 2050.
Adaptive Capacity: Research indicates that while 70% of workers in AI-exposed roles have the capacity to transition, roughly 6.1 million workers—86% of whom are women—lack this adaptive capacity.
Monetary Policy and Financial Stability
Costs of Bank Regulation: Tighter-than-expected bank regulation has been found to reduce lending and raise unemployment for several years, suggesting higher short-run macroeconomic costs than previously estimated.
Housing Market Stabilization: Contractionary monetary policy is found to be at least twice as effective at mitigating housing market overvaluation as expansionary policy is at exacerbating it.
Global Shocks: Global shocks now play a much larger role in driving interest rate movements than in previous decades, with their contribution doubling since 1999.
Trade and Global Competition
Tariff Implications: Recent Supreme Court rulings striking down certain tariff authorities have left businesses to reassess trade paths. Experts note that tariffs on intermediate goods (inputs) generally have more severe economic effects than those on final consumer goods.
The "China Shock": Research suggests China's currency peg amplified the "China shock" between 2000 and 2012; a floating exchange rate might have eliminated 59% of the resulting U.S. manufacturing decline.
USMCA Review: The U.S.-Mexico-Canada Agreement is undergoing its first joint review to determine whether to renew, revise, or terminate the pact.
Productivity and Labor Supply
Manufacturing Quality: Official measures may be understating manufacturing productivity growth by failing to sufficiently adjust for quality improvements, particularly in high-tech sectors.
Medicaid and Work Hours: The decline in average hours worked in the U.S. relative to other advanced economies is partially attributed to an increase in publicly provided health benefits like Medicaid for non-employed individuals.
Apprenticeship Barriers: High administrative burdens and inconsistent standards continue to deter U.S. employers from launching needed apprenticeship programs.
Next Step: Would you like me to draft a specific introductory paragraph for this round-up or perhaps format it as a ready-to-send Gmail draft?
THE BROOKINS INSTITUTE
The Hutchins Roundup
Posted April 16, 2026 Reposted April 17, 2026
Economists Predict Large, but Not Unprecedented, AI Effects
Economists and the public were surveyed on short- and long-run projections of economic growth and labor market outcomes under various AI growth scenarios.
Summary: In a "rapid AI" scenario where AI outperforms humans in most cognitive and physical tasks, real GDP growth could rise to 3.5% by 2050. However, this would likely be accompanied by pre-World War II levels of inequality, with the top 10% of households holding 80% of the wealth.
Key Takeaway: While AI is expected to significantly boost growth, the primary economic challenge will be managing a projected 55% labor force participation rate and severe job loss among blue-collar workers.
Economic Costs of Bank Regulation
University of Maryland researchers Thomas Drechsel and Ko Miura identified how regulatory news affects bank lending and the broader economy.
Summary: Tighter-than-expected bank regulation reduces bank lending and increases the unemployment rate for several years. A regulatory shock that lowers bank stock prices by 1% is estimated to reduce lending by 0.5% to 1% and raise unemployment by roughly 10 basis points over two to three years.
Key Takeaway: The short-run macroeconomic costs of bank regulation are higher than previously estimated, though they may still be justified by long-run financial stability benefits.
Manufacturing Productivity Decline Explained by Quality Gains
Enghin Atalay of the Federal Reserve Bank of Philadelphia argues that official measures have failed to account for product quality improvements.
Summary: Official statistics understate annual productivity growth by 5.5 percentage points for computer and electronics manufacturing and 0.35 percentage points for manufacturing overall. These adjustments reverse the reported decline in manufacturing productivity between 2009 and 2023, bringing it in line with the rest of the economy.
Key Takeaway: Reported declines in manufacturing productivity are likely a measurement error resulting from insufficient adjustment for rapid innovation in high-tech industries.
Quote of the Week
"I don’t believe this is going to get embedded into inflation expectations. We can see forward inflation expectations; they are well-anchored... I think the Fed is doing the right thing by sitting and watching." — Scott Bessent, Treasury Secretary