The National Academies

11/21/2024 | Press release | Distributed by Public on 11/21/2024 10:17

U.S. Should Build Capacity to Rapidly Detect and Respond to AI Developments; New Report Identifies Workforce Challenges and Opportunities

Share

U.S. Should Build Capacity to Rapidly Detect and Respond to AI Developments; New Report Identifies Workforce Challenges and Opportunities

News Release| November 21, 2024
WASHINGTON - It is imperative to improve near real-time observation and tracking of progress in artificial intelligence (AI), its adoption, and its impacts on the workforce, and to widely share this information to better inform and equip workers, policymakers, and others who will have to flexibly respond to AI developments, says a new report from the National Academies of Sciences, Engineering, and Medicine.
With rapid technical progress in AI likely to continue, the precise trajectories of AI-enabled futures are uncertain, underscoring the need for society to be able to understand and respond nimbly to new developments, the report says. AI's progression, its labor market consequences, and its effects on society will depend on both the rate and direction of AI's capabilities and on demographic, social, institutional, and political forces.
AI has emerged as a general-purpose technology with sweeping implications, the report says. Innovations in neural networks and a shift to larger, unannotated datasets - often aggregated from the internet - have largely driven the recent surge of technical progress in AI. Adoption has been facilitated by the availability of cloud services and ease of integration into computing platforms. Dramatic advances with generative AI systems have been accompanied by steady progress in other areas of AI such as robotics, with self-driving vehicles being tested in multiple cities, for example, and in other forms of machine learning.
The report, requested by Congress as part of the 2021 National Defense Authorization Act, examines how AI is likely to complement or replace human labor, reshape job markets, and influence workforce dynamics. It emphasizes that AI should be developed in accordance with society's shared values and goals, and it should be used to enhance human well-being, augment collective capabilities, and support a well-equipped future workforce. It also lays out scenarios to help guide research and policymaking toward beneficial deployment of AI, as it will require intentional design to deploy AI to enhance human labor, complement expertise, and create new forms of valuable work.
"AI is a powerful tool, but it is still a tool to be directed by humans. Collectively, we - as citizens, businesses, nonprofits, researchers, workers, and government - have agency over the type of society we want to be a part of and how we choose to use AI," said the report's study committee co-chair, Erik Brynjolfsson, director of the Stanford Digital Economy Lab and Jerry Yang and Akiko Yamazaki Professor at the Stanford Institute for Human-Centered AI. "Our ability to make informed choices, to prepare for the changes ahead, and to build flexible strategies and policies requires major improvements to our capacity to sense, analyze, and respond to AI-related developments rapidly."

Measurement Issues

Collecting and transparently sharing information on changes in AI, its adoption, demand for various expertise, and workforce impacts can support workers in making career and continuing education decisions during a time of rapid change, the report says. Government leaders and others can have positive influence by improving data collection efforts, including high-frequency, real-time tracking of AI use by businesses and workers. The availability of data that track the use and impacts of advanced technologies has increased since a National Academies assessment in 2017, but knowledge gaps remain, and data access and integration are ongoing challenges. Opportunities for improvement include:

  • Creating new public-private data partnerships to widely share real-time data on skills supply and demand, wages, and continuing education opportunities
  • Measuring impacts on AI adoption across and within economic sectors and geographic regions
  • Measuring impacts of new technologies on marginalized groups and communities
  • Exploring development of an independent, nonprofit, government-chartered entity to support public-private data sharing and integrated analysis

AI and Productivity

Substantial improvements in AI's capabilities, along with its broad applicability to cognitive tasks and ability to spur complementary innovations, offer the promise of significant improvements to productivity. The report emphasizes that achieving the full benefits of AI will require investments and research into new skills and organizational processes and structures.
Effects across the economy will not be equal because there are important differences in the exposure of different sectors and occupations to AI, the report notes. AI adoption in most companies is still relatively low, but the technology has increasingly permeated economic activity in certain areas, such as manufacturing, advanced driver assistance systems in cars, customer service, and e-commerce.
The benefits of AI's productivity gains may not accrue inclusively, resulting in greater income and wealth inequality, the report says. Evidence from the information, communications, and internet technological revolutions indicates reason for concern, as real wages in the U.S. and other advanced economies grew more slowly than labor productivity, resulting in a decoupling of productivity and wage growth.
"We should anticipate productivity gains across a wide cross section of the economy, and we are already seeing those gains in certain areas such as software development," said committee co-chair Tom Mitchell, Founders University Professor at Carnegie Mellon University and founder of the world's first machine-learning department. "But many of these gains will require that workers learn to use new AI tools effectively. Providing workers with retraining opportunities and keeping them informed of shifting demands for different skills will be essential to assuring that the workforce benefits from advancing AI technology, and that the new wealth generated from these advances is shared widely."
Estimates of productivity gains also rely on the assumption that displaced workers find jobs of comparable productivity, which would require large government and institutional investments in worker training, skill development, and transition support programs. The report also details how AI comes with considerable societal risks, including threats to privacy, the potential for discrimination and bias, risks to political stability, and national security concerns - and how research is needed to deal with these and the social implications of emerging AI technologies. Overall, adoption of AI will alter the nature of many jobs, and widespread adoption may catalyze fundamental changes in the structure of jobs and industries.

Workforce Implications

The most relevant concern for workers is how AI impacts the labor market by reshaping the demand for expertise - whether it increases the value of skills and expertise that people possess or erodes that value by providing cheaper substitutes. The report cautions that advances in AI could apply downward pressure on wages and heighten risks of worker surveillance, privacy concerns, and creative output ownership issues.
AI is likely to speed the rate of automation of "mass expertise" for tasks such as taking retail inventory at stores and stocking warehouses. It is also likely to enable the partial automation of intellectual services provided by "elite expertise" in an array of currently high-value fields - for example, writing legal documents, a task typically completed by lawyers or legal professionals.

It is more challenging to anticipate what types of worker expertise will be augmented by AI and what new forms of expertise will be demanded. However, by providing relevant information, guidance, and potentially digital guardrails, AI could enable workers with expert judgment to perform a broader range of expert tasks - including the vast number of workers whose jobs require judgment, problem-solving, and decision-making, such as teachers, health technicians, software developers, customer support workers, and repair people. Policymakers could help by supporting research into how best to design human-complementary AI, how the cost of capital and labor affect business decisions on AI, and how to provide individuals with control and compensation for creative work.

Effects on Education and Training

AI will impact both the overall supply and demand for education, as well as how it is supplied, the report says. It has the potential to create positive transformations in the education system, improving learning outcomes by making education more personalized, engaging, and cost-effective. While the last decade has seen a rise in online educational platforms - which often apply machine learning to understand the effectiveness of teaching materials - integrating AI could improve their quality and personalization. Large language models could also effectively function as tutors, helping students as they write essays and solve math problems. But research and careful experimentation are needed to determine best practices and how and whether curricula should be shifted, in order to best prepare the next generation with the knowledge and tools needed to leverage AI in their everyday lives and in their future careers.
Increased demand for retraining and other forms of continuing education is expected, as workers navigate the changing landscape of job opportunities and demand shifts for different skills and expertise. Changes to worker training policies and practices might improve workers' ability to make use of educational opportunities, and the report highlights ways to assist workers and foster workplace flexibility.
The report details how public and private investments to increase universal access to things like high-speed internet will be necessary to reap the potential benefits of AI in education, and how additional efforts to incorporate safeguards, test the effectiveness of specific use cases, train teachers, and take advantage of computer-assisted learning technologies will also be needed. While there are fears that AI may disrupt education in various ways, appropriate research, investments, and applications of AI and AI-enhanced technologies could be used to facilitate learning and help flatten educational inequalities.

AI Opportunities for Government

The report says decision-makers need to create policies that will be robust in the face of various possible technology advances and future timetables. Some of the opportunities for government leaders to influence the direction, robustness, and speed of AI development include supporting:

  • Enhancement of AI expertise within the federal government to help manage effective investment oversight, and regulation across all mission areas, from transportation to national security
  • Basic research in AI and research into standards and guardrails
  • Research toward applications of high societal priority, such as education and training
  • Incentives, standards, and regulations to encourage transparency, privacy protections, and a level playing field
  • Initiatives like the National AI Research Resource and the Microelectronics Commons to provide hubs for computational resources and the talent needed to keep U.S. universities at the forefront of AI research and safety developments

Core Technology

While rapid technical progress in AI and the changes it catalyzes are likely to continue in the near term, there are a range of factors that could alter the speed and trajectory of advances, and the report identifies both drivers and inhibitors for AI's future technical progress and adoption.
AI systems today remain imperfect, subject to error, manipulation to produce false results, "hallucinations," and biased behavior - all of which are likely to persist to some degree. The need to address these shortcomings; the prerequisite that technology be socially acceptable and trusted by the public; privacy concerns; diminishing opportunities in the largest AI research advances for universities; and other forces can slow the rates of technical advancement and societal adoption, the report says. AI adoption may also not proceed as rapidly as many expect due to legal and regulatory constraints, lack of consumer trust, and cybersecurity risks. Upgrades in computational power, algorithmic innovation, increasing investment, and the volume of online data for training are all potential drivers of future technical progress.
The study - undertaken by the Committee on Automation and the U.S. Workforce: An Update - was sponsored by the National Science Foundation.
The National Academies of Sciences, Engineering, and Medicine are private, nonprofit institutions that provide independent, objective analysis and advice to the nation to solve complex problems and inform public policy decisions related to science, engineering, and medicine. They operate under an 1863 congressional charter to the National Academy of Sciences, signed by President Lincoln.

Contact:

Josh Blatt, Media Relations Officer
Office of News and Public Information
202-334-2138; email [email protected]

Featured Publication

2024

Artificial Intelligence and the Future of Work

Advances in artificial intelligence (AI) promise to improve productivity significantly, but there are many questions about how AI could affect jobs and workers.

Recent technical innovations have driven the rapid development of generative AI systems, which produce text, images, or other content based on user requests - advances which have the potential to complement or replace human labor in specific tasks, and to reshape demand for certain types of expertise in the labor market.

Artificial Intelligence and the Future of Work evaluates recent advances in AI technology and their implications for economic productivity, the workforce, and education in the United States. The report notes that AI is a tool with the potential to enhance human labor and create new forms of valuable work - but this is not an inevitable outcome. Tracking progress in AI and its impacts on the workforce will be critical to helping inform and equip workers and policymakers to flexibly respond to AI developments.

Read Full Description

Recent News

U.S. Winners of 2024 Nobel Prizes Honored

Scaling Up Innovations in STEM Education

Fellows Program Celebrates 50 Years

Science Is Neither Red Nor Blue

  1. Load More...