A report spearheaded by Americans for Responsible Innovation and the University of Notre Dame offers recommendations in areas ranging from data and research needs to workforce development, safety nets and industry policy, for addressing the labor market opportunities and “considerable displacement risks” posed by artificial intelligence.
“Although AI policy discourse has largely centered on technological innovation, market dynamics, and global competition, concrete strategies for labor market disruption remain under developed. This white paper aims to fill this gap, emphasizing pragmatic policies implementable at federal and state levels and adaptable to international contexts,” according to the report released Sept. 2 and shared exclusively beforehand with Inside AI Policy.
The 73-page report, “Proactively Developing & Assisting the Workforce in the Age of AI,” springs from “a workshop on AI and labor policy, organized by the Keough School of Global Affairs and the Institute for Ethics and the Common Good at the University of Notre Dame, held in Washington, D.C., on March 11, 2025. The paper is co-branded and distributed in collaboration with Americans for Responsible Innovation,” according to the executive summary.
The report assesses current AI capabilities and looks at “predictive machine learning” applications across industry sectors while also considering the quite different batch of use cases specifically for generative AI, and the types of occupations affected by each type of AI.
“Machine learning algorithms are used to analyze data, forecast outcomes, and support decision-making in areas such as healthcare, finance, retail, and manufacturing, among others,” it says.
“Generative AI systems,” on the other hand, “built on LLMs, now produce essays, computer code, marketing reports, and more by predicting the most likely next word in a sequence based on learned language patterns as well as information searched online.”
The report says, “As these models acquire multimodal capabilities -- handling text, images, audio, and video -- they are referred to as generative AI more broadly. They can now perform text-to-image, text-to-video, and even image-to-text and video-to-text tasks, expanding the boundaries of automation to include creative and interpretive tasks.”
Next, the report says, “A key emergent AI capability is agentic AI, that is systems capable of autonomously performing multi step tasks by integrating reasoning, planning, and execution. Like some forms of traditional software, AI agents can execute tasks in sequence. Unlike traditional software, agents ‘independently’ determine how to achieve goals using advanced reasoning capabilities provided by large language models. Recently coding agents have been able to perform the full spectrum of coding tasks and develop and launch software applications with minimal human involvement.”
The report raises issues around artificial general intelligence and also considers “physical AI” and advanced robotics.
All of these use cases have impacts on different aspects of the labor market, the report explains.
“Policies designed to assist workers in the age of AI should be grounded in empirical evidence and fundamentally ensure human dignity. In this section, we review the current state of research on AI’s impact on labor markets,” the report says. “First, we discuss conceptual frameworks and approaches developed to understand and examine AI’s effects on labor and productivity.”
“Finally,” it says, “we present case studies that provide deeper insights and nuanced perspectives on AI’s impact on workers. Given that AI’s influence on labor will likely continue evolving in the coming years, these conceptual frameworks, empirical findings, and case studies will also evolve. Policymakers should continue to examine and incorporate this evidence into their decisions.”
The report says “key insights regarding the conceptual frameworks for empirical analysis, job displacement versus creation, productivity and wage outcomes, and industry-specific differences” include:
The task-based framework and AI exposure literature shows that AI affects work primarily at the task level, given that occupations consist of tasks with different degrees of susceptibility to automation. Exposure scores provide estimates of potential AI impacts, but actual adoption depends significantly on economic and organizational factors. Highly educated workers experience higher exposure yet are also more likely to benefit from AI adoption. Ultimately, labor market outcomes depend on actual AI adoption rather than merely exposure.
Further, studies show wide variance in labor market impacts and “productivity and wage effects.”
Data, research and measurement
On policy options addressing “data, research, and measurement,” the report says, “Policymakers are confronted with substantial uncertainty due to the rapid advancements in AI technologies, which could significantly affect labor markets and the wider economy. Policymakers must effectively measure and comprehend these developments to ensure timely responses.”
It says, “Key actions include modernizing the federal statistical system, enhancing real-time data collection, and promoting robust institutional collaboration to accurately track the evolving impacts of AI.”
“Enhanced data collection and analysis,” it says, “will facilitate better forecasting, targeted interventions, and informed policymaking, ensuring agility and responsiveness to AI-driven shifts.”
It offers a series of “enhancements to existing federal data programs that would substantially improve our ability to monitor how AI is reshaping work.”
“Ultimately, updating classification systems and federal data collection efforts to reflect AI-driven changes is not merely a technical issue,” the report says. “It is essential infrastructure for understanding the future of work, identifying emerging risks, and enabling timely, evidence-based policymaking. Integrating AI-relevant metrics would allow more precise identification of vulnerable jobs, emerging skill gaps, and potential pathways for worker transitions.”
The report also recommends “four high-impact opportunities for collecting more detailed, longitudinal, and task-level information at the worker level.”
“To effectively address the rapid and unpredictable impacts of AI on labor markets, it is imperative for policymakers to prioritize comprehensive data collection, robust research, and strategic measurement methods,” the report says.
“Innovations such as real-time adaptive occupation mapping, enhanced data products from the [Bureau of Labor Statistics], detailed worker-level data, and leveraging private-sector insights will significantly strengthen understanding of AI’s effects. Furthermore, investing in integrated and interoperable data systems, enhancing longitudinal research, and expanding institutional capacities will improve evidence-based policymaking.”
The report says, “Strategic use of AI itself within data collection processes can improve operational efficiency, accuracy, and responsiveness. Together, the recommendations in this section can equip policymakers and researchers with the necessary tools to anticipate workforce shifts, effectively target interventions, and manage economic transitions with agility and precision in the face of AI-driven labor market disruptions.”