The Wired Belts are the new Rust Belts

AI’s next target will be regions with America’s highest-paid jobs

Tufts University

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

New Fletcher/Tufts index maps AI job displacement risk across 784 U.S. occupations, 530 metro areas, and 50 states—9.3M jobs at risk, $757B in income.

view more 

Credit: Anthropic Economic Index (2025); Tomlinson et al. (2025): Digital Planet, The Fletcher School, Tufts University

MEDFORD, MA — March 24, 2026 — Digital Planet, the research center at the forefront of researching the AI transformation at The Fletcher School at Tufts University, today released the American AI Jobs Risk Index. It is a first-of-its-kind data-driven framework that maps the potential of AI-driven job vulnerability across every major occupation, industry, metropolitan area, and state in the United States. Drawing on 15 years of labor market data and the most current AI adoption research, the Index goes beyond prior studies by measuring actual vulnerability to job loss — not merely exposure — and connecting that vulnerability directly to projected income loss and geography.

The Index projects approximately 9.3 million U.S. jobs are at risk of displacement in the next 2–5 years, with a plausible range of 2.7 to 19.5 million depending on alternative adoption scenarios. Associated household income at risk spans $200 billion to $1.5 trillion annually, with a midpoint of roughly $757 billion — equivalent to the economies of Belgium and, under faster AI adoption, approaching that of South Korea. The assessment does not incorporate job creation effects given lack of robust available data.

"We already know that AI is not just automating routine tasks — it is moving up, targeting the cognitive and analytical work that defines high-skill, high-wage careers. The jobs of the future will be secured by those with a combination of subject-matter expertise, critical-thinking skills for human judgment, and knowledge of AI and how to use it. Our index makes clear that the question is no longer whether AI will displace significant numbers of workers, but in which states and cities, how fast, and whether we are prepared by taking pre-emptive action. The geography of this disruption has real political consequences: the states and metros most at risk are already the most active in seeking AI regulation — and the federal government is telling them to stand down. That collision will define the economic and political landscape of the next decade."

— Bhaskar Chakravorti, Dean of Global Business, The Fletcher School at Tufts University, and Chair of Digital Planet

Key Findings

The Scale of Risk Is Historic. Expected American job loss amounts to a wipeout equivalent to the economy of Belgium. Industry-wide vulnerability averages approximately 6%, but the steepest risks sit in Information (18%), Finance and Insurance (16%), and Professional, Scientific, and Technical Services (16%). High-earning knowledge workers — Writers and Authors (57%), Computer Programmers (55%), and Web and Digital Interface Designers (55%) — face the highest rates of job displacement by occupation.

$757 Billion in Annual Income Is on the Line. Occupations vulnerable to AI-driven displacement account for an estimated $757 billion in annual U.S. wages and salaries. The greatest absolute income losses fall on Software Developers, Management Analysts, and Market Research Analysts, reflecting both high salaries and large worker populations.

4.9 Million Workers Are at a Tipping Point. The Index identifies 33 "tipping point" occupations — spanning 4.9 million workers — that swing from under 10% to over 40% displacement risk depending on how quickly AI adoption accelerates. Ironically, over one million workers whose jobs involve studying, building, or reporting on AI itself face 26–55% displacement rates. AI is rewriting how we write history: Historians top the automation ranking, with 67% of their tasks expected to be automated.

Augmentation Is a Displacement Pipeline. The more AI enhances worker efficiency, the more expendable individual workers become. The Index finds a strong relationship between a job's augmentation potential and its displacement risk. For every 1 percentage point increase in job automation, the Index projects a 0.75 percentage point job loss.

Innovation Hubs Are the Most Exposed. Major urban centers — New York, Los Angeles, Washington D.C., San Francisco, Chicago, Dallas, and Boston — each face at least $20 billion in projected annual income losses. Silicon Valley (San Jose metro) leads all regions with 9.9% of jobs at risk. University towns including Durham-Chapel Hill, Boulder, Ann Arbor, Ithaca, and Madison rank among the top 25 most vulnerable metros — the "Wired Belts" are the first to rust.

The Safe Zone Is the Near-Poverty Zone. 38% of American workers are AI-proof — but these are also the country's lowest-paid jobs. Physical, manual, and variable-condition work (roofers, orderlies, dishwashers) face less than 1% displacement. The occupations AI cannot touch are largely those the economy has always undervalued.

State-Level Risk Carries Political Consequences. D.C. (11.3% vulnerability), Massachusetts, Virginia, Maryland, Washington, and Colorado face the highest proportional exposure. California, Texas, New York, Florida, and Illinois will experience the largest absolute losses. Critically, states most at risk from AI job displacement are legislating 4× more on AI than the safest states — ‎ and the federal government, via a December 2025 Executive Order, has directed the Justice Department to challenge state-level AI laws and threatened to withhold Broadband Equity, Access, and Deployment (BEAD) funding from states that proceed.

What Sets This Index Apart

The American AI Jobs Risk Index enters a field that includes analyses from Goldman Sachs, MIT's Iceberg Index, the Yale Budget Lab, Stanford's "Canaries in the Coal Mine," and the WEF Future of Jobs Report. It breaks new ground in four areas:

• Vulnerability over exposure — the Index measures the likelihood that AI exposure translates into actual job loss, not just the theoretical reach of AI into a role's task mix.
• Geographic granularity — every metro area and non-metro region in the country is ranked, not just national-level occupational exposure.
• Economic impact — vulnerability findings are directly linked to projected income loss, giving policymakers a clear picture of dollar-denominated stakes.
• Policy and local action orientation — findings translate directly into recommendations for action where the pain will be felt and the leverage points exist — at the local level in states and cities — across policymakers, businesses, technologists, and civil society.

Recommendations FOR ACTION

The Index yields a forward-looking assessment of AI’s potential impact on employment with actionable recommendations.

Policymakers are urged to modernize unemployment insurance to cover gig and contract workers most exposed to automation, introduce wage insurance for AI-displaced workers re-employed at lower wages, and create state-level AI Workforce Transformation Groups. They should also create a unified AI labor market measurement initiative and require companies to disclose AI workforce data. State and local policymakers should aim for regulations and safeguards with their constituencies’ interests in mind.

Businesses are called on to invest in targeted reskilling and internal mobility programs and to track and publish data on AI's actual workforce impact. They should also approach AI deployment as augmentation first, not replacement, and decisions directly affecting workers (hiring, scheduling, performance evaluation) should retain meaningful human review.

Technologists should partner with educators to build stackable AI credentials and lifelong learning infrastructure, while educators should prepare students with a combination of AI literacy, subject-matter expertise, and critical-thinking skills.

Civil society organizations are encouraged to independently monitor displacement, connect economic disruption to political action, and advocate for education reform targeted to specific at-risk roles.

 

About the American AI Jobs Risk Index

The American AI Jobs Risk Index was researched and authored by Digital Planet at The Fletcher School at Tufts University. The Index ranks 784 occupations, 530 metropolitan and non-metropolitan areas, 50 U.S. states, and 20 industry sectors by vulnerability to AI-driven job displacement. The Index will be updated as AI capabilities and labor market conditions evolve.

Full report and methodology: https://digitalplanet.tufts.edu/ai-and-the-emerging-geography-of-american-job-risk-page/ 

About Digital Planet at The Fletcher School, Tufts University

Digital Planet, an interdisciplinary research initiative of The Fletcher School’s Institute for Business in the Global Context, is dedicated to understanding the impact of digital innovations on the world and providing actionable insights for policymakers, businesses, investors, and innovators. The Fletcher School, founded in 1933, is a premier graduate school of international affairs, preparing leaders across government, business, and civil society to advance solutions to the world’s most pressing challenges.

Follow Fletcher on X (@FletcherSchool), Instagram (@thefletcherschool), Facebook (@fletcherschool), and LinkedIn. Visit fletcher.tufts.edu for news and events.

 

Generative AI-powered forecasting for sustainable urban development

A memory-aware multi-conditional generative model for predicting future urban layouts

Japan Advanced Institute of Science and Technology

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

Overview of the Memory-aware Multi-Conditional generation Network (MMCN) framework for forecasting future urban layouts. The system integrates multiple modules, including a spatial memory module that captures contextual information from neighboring regions, a multi-prompt fusion module that combines urban condition inputs such as building density, building height, and road networks, and a multi-conditional control module that guides a diffusion-based generative model. Together, these components enable the model to generate spatially coherent urban layout predictions while maintaining continuity across adjacent areas.

view more 

Credit: Associate Professor Haoran Xie from JAIST

Researchers introduce a novel generative AI-driven framework, MMCN (Memory-aware Multi-Conditional generation Network), for forecasting future urban layouts by jointly considering building density, building height, transportation networks, and historical development patterns. Leveraging a generative architecture-enhanced diffusion model with multi-conditional control, semantic prompt fusion, and spatial memory embedding, MMCN offers a novel approach to modeling complex urban evolution. This framework provides a powerful tool to explore sustainable urban development, demonstrating AI’s transformative potential in urban design.

Environmental sustainability in urbanization has become a critical global concern as cities expand at unprecedented rates. Urban design faces the challenge of making long-term decisions about infrastructure, building development, transportation networks, and land use, all of which shape the future structure and sustainability of cities. These decisions are inherently complex, as urban growth emerges from the interaction of multiple factors, including building density, building height, road networks, and historical development patterns, which evolve together over time. Traditional urban design methods often struggle to capture these interconnected dynamics, making accurate forecasting of urban development impossible.

In response to this challenge, artificial intelligence (AI) has emerged as a promising tool for modeling complex spatial patterns and supporting data-driven urban planning. Yet, many existing generative AI-based models produce fragmented predictions because they may have difficulty in effectively integrating multiple urban development factors or maintaining spatial continuity across large areas.

To address these limitations, researchers at the Japan Advanced Institute of Science and Technology (JAIST) and Waseda University, Japan, developed a novel AI-driven framework called the Memory-aware Multi-Conditional generation Network (MMCN). The research team was led by Associate Professor Haoran Xie (JAIST and Waseda University) and included Doctoral Student Xusheng Du from JAIST and Professor Zhen Xu from Tianjin University, China, among others. Their study was published online on March 2, 2026, and will be published in Volume 141 of the top journal in urban design, Sustainable Cities and Society, on May 1, 2026.

. Explaining the motivation behind the study, Dr. Xie said, “We aimed to bridge the gap between current AI capabilities and the practical needs of urban planners by developing a predictive model capable of forecasting future urban layouts while simultaneously considering multiple urban development factors and historical evolution patterns, as inspired by the actual decision-making workflow from professional planners.”

The MMCN model relies on multi-temporal spatial data, including building layouts, building density, building height, and transportation networks, which were standardized into 512 × 512-pixel patches for model training. Especially, this model adopted the urban layout data of Shenzhen due to it being the most rapidly developing city in China. The network architecture combines a diffusion model with a multi-conditional control mechanism, allowing diverse urban factors to guide the generation process. A semantic prompt fusion module encodes information from each input type, while a spatial memory embedding component preserves contextual information from neighboring regions, ensuring continuity across patches. Multiple conditional generation branches integrated with the diffusion model form the core generative model, enabling the production of realistic, coherent urban layouts that remain consistent with historical patterns. Data training uses denoising and edge-stitching loss functions to enhance reconstruction accuracy and smooth transitions across patch boundaries. This approach allows MMCN to model complex interactions among urban variables and generate spatially consistent forecasts of urban development.

Experimental results demonstrated the framework’s effectiveness. MMCN outperformed baseline methods such as Pix2Pix, CycleGAN, and Instruct-Pix2Pix, achieving a Structural Similarity Index (SSIM) of 0.885 and a Boundary Intersection over Union (IoU) of 0.642, indicating strong structural fidelity and spatial continuity. Qualitative analysis further confirmed that MMCN generates realistic, coherent urban layouts with continuous road networks and well-organized building clusters, whereas baseline models often produce fragmented roads, duplicated structures, or disconnected patterns. These findings highlight the importance of combining multi-factor conditioning, spatial memory mechanisms, and learning from historical patterns within a unified generative framework. Additional cross-city experiments using data from Shanghai and Tianjin in China further demonstrated the model’s ability to produce stable and consistent urban layout predictions under diverse spatial conditions.

Beyond technical performance, MMCN offers practical benefits for urban design. By simulating potential growth scenarios, the framework allows planners to evaluate the long-term consequences of development strategies, supporting more informed and sustainable decisions. This aligns with the Sustainable Development Goals, particularly those focused on creating resilient and inclusive cities.

Looking ahead, the researchers envision several enhancements. Integrating climate models could enable assessment of environmental impacts, while including socio-economic data, could support more comprehensive forecasts. “Interactive planning tools built on MMCN could facilitate community and stakeholder engagement in urban design, promoting collaborative planning,” said Dr. Xie. He added, “Expanding the dataset to include cities with diverse morphologies would improve the model’s generalizability, making it applicable across different urban contexts worldwide.”

In conclusion, MMCN represents a significant advancement in AI-assisted urban design, offering a novel approach to forecasting urban layout evolution by integrating multiple spatial factors and historical patterns. By producing accurate, spatially coherent predictions, it provides a powerful tool for guiding cities toward more resilient, livable, and sustainable futures in an increasingly urbanized world.

***

 

Illustration of how MMCN predicts future urban layout patches. The model receives multiple inputs, including neighboring urban patches, historical building layouts, density maps, height maps, and road networks, which are used as prompts and control conditions for a diffusion-based model. Using these inputs, the model can generate a layout for a target region and stitches the generated patch together with surrounding areas to produce a spatially consistent urban map.

Credit

Associate Professor Haoran Xie from JAIST

Reference

Authors: Xusheng Du, Chengyuan Li, Qingpeng Li, Yuxin Lu, Yimeng Xu, Ye Zhang, Zhen Xu, and Haoran Xie

DOI: https://doi.org/10.1016/j.scs.2026.107272

 

About Japan Advanced Institute of Science and Technology, Japan

Founded in 1990 in Ishikawa prefecture, the Japan Advanced Institute of Science and Technology (JAIST) was the first independent national graduate university that has its own campus in Japan. Now, after 30 years of steady progress, JAIST has become one of Japan’s top-ranking universities. JAIST strives to foster capable leaders through a state-of-the-art education system that emphasizes diversity; about 40% of its alumni are international students. The university has a unique style of graduate education based on a carefully designed, coursework-oriented curriculum to ensure that its students have a solid foundation on which to conduct cutting-edge research. JAIST also works closely with both local and overseas communities by promoting industry–academia collaborative research.  

Website: https://www.jaist.ac.jp/english/

About Associate Professor Haoran Xie from Japan Advanced Institute of Science and Technology, Japan

Dr. Haoran Xie is an Associate Professor at the Japan Advanced Institute of Science and Technology (JAIST) and Waseda University. At JAIST, he directs the Human-Centered AI Laboratory. His research focuses on human-centered generative AI, exploring how emerging technologies can enhance human capabilities through interactive computer graphics, deep learning, and human–computer interaction. His work spans creative applications, including generative design of anime, architecture, and fashion, as well as physical intelligence and robotic learning.

Funding information

This work was supported by JST SPRING, Japan Grant Number JPMJSP2102, JST BOOST Program Japan Grant Number JPMJBY24D6, the National Key R&D Program of China under Grant 2024YFC3808104-01, and the National Natural Science Foundation of China under Grant 52508023.

---

Journal

Sustainable Cities and Society

DOI

10.1016/j.scs.2026.107272 

Method of Research

Computational simulation/modeling

Subject of Research

Not applicable

Article Title

AI-driven urban evolution forecasting: A unified memory-aware multi-conditional generation framework for sustainable development planning

Article Publication Date

1-May-2026

 

JMIR Publications analyzes the gap between AI law and patient reality in health care

JMIR Publications

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

Anshu Ankolekar, PhD., JMIR Correspondent

view more 

Credit: Anshu Ankolekar

(Toronto, March 23, 2026) JMIR Publications today announced the release of a timely new article in its News and Perspectives section, examining the legal and ethical complexities of the right to explanation for patients in the era of artificial intelligence. The article, "The Right to Understand in Health Care AI," explores a critical tension: while the European Union’s AI Act provides a legal basis for transparency, the technical and clinical reality of meaningful explanations remains largely undefined.

Authored by Anshu Ankolekar, JMIR Correspondent, the report highlights that as high-risk AI systems become standard in medical imaging and diagnostics, patients are increasingly entitled to ask, "Why did the computer conclude this?" However, the opacity of advanced algorithms often leaves even the most experienced clinicians unable to provide an answer that is both technically accurate and practically useful for the patient.

The Paradox of Clinical AI Transparency

The analysis points to several significant hurdles that prevent current legal frameworks, such as the EU AI Act and GDPR, from translating into better patient care:

• The Interpretability Trade-off: The most accurate AI models often operate through millions of parameters that are impossible for humans to trace fully. Forcing simpler, more explainable models could potentially sacrifice diagnostic accuracy, creating a direct conflict with patient safety.

• Automation Bias: Research indicates that incorrect AI suggestions can pull clinicians toward an incorrect diagnosis regardless of experience level. An explanation delivered by a clinician who has already deferred to an algorithm may not reflect an independent clinical assessment.

• The Literacy Barrier: Between 22% and 58% of EU citizens report difficulty understanding health information. Providing technical detail on algorithmic logic often leads to cognitive overload rather than informed consent.

Shifting from Compliance to Effectiveness

The article argues for a paradigm shift: moving away from a check-the-box compliance approach toward one focused on decision-relevant clarity. Experts suggest that a truly useful patient-facing explanation must address what the system recommends, how confident it is, and what the known performance gaps are for specific populations.

To bridge this gap, the report calls for:

• Co-design Partnerships: Developers must test explanation systems with actual patients and advocates to ensure they meet real-world needs.

• Institutional Support: Health care systems need to allocate specific time for AI discussions and train staff to navigate these complex conversations.

• Standards for Comprehension: Policy makers should prioritize digital health literacy and develop standards that measure whether a patient can actually use the information provided to make a choice.

"The EU AI Act provides the legal foundation, but the capacity to deliver an explanation that a patient can genuinely use is shaped by forces the law alone cannot govern," the report concludes. "What patients need now are answers they can use."

Please cite as:

Ankolekar A. The Right to Understand in Health Care AI. J Med Internet Res 2026;28:e95090

URL: https://www.jmir.org/2026/1/e95090

DOI: 10.2196/95090

 

About JMIR Publications News and Perspectives

JMIR Publications is a leading open access publisher of digital health research. The News and Perspectives section is the newest addition to its portfolio, established to bring the rigor and integrity of academic publishing to scientific journalism. The section features well-researched, expert-driven content from the Scientific News Editor, Kayleigh-Ann Clegg, PhD, and a network of specialist JMIR Publications Correspondents to keep the digital health community informed, inspired, and ahead of the curve.

About JMIR Publications

JMIR Publications is a leading open access publisher of digital health research and a champion of open science. With a focus on author advocacy and research amplification, JMIR Publications apartners with researchers to advance their careers and maximize the impact of their work. As a technology organization with publishing at its core, we provide innovative tools and resources that go beyond traditional publishing, supporting researchers at every step of the dissemination process. Our portfolio features a range of peer-reviewed journals, including the renowned Journal of Medical Internet Research. 

To find out more about JMIR Publications, visit jmirpublications.com or connect with them on Bluesky, X, LinkedIn, YouTube, Facebook, and Instagram.

Media Contact:

Dennis O’Brien, Vice President, Communications & Partnerships

JMIR Publications

communications@jmir.org

+1 416-583-2040

The content of this communication is licensed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, published by JMIR Publications, is properly cited.

---

Journal

Journal of Medical Internet Research

DOI

10.2196/95090 

Method of Research

Commentary/editorial

Subject of Research

People

Article Title

The Right to Understand in Health Care AI

Article Publication Date

19-Mar-2026