On June 7, 2023, Aspen Digital convened a roundtable of 21 experts in economics, technology, human resources, labor, and policy to discuss the labor implications and opportunities of artificial intelligence (AI). The conversation spoke to risks both individual and systemic but also acknowledged potential positive impacts from AI and glimmers of hope on the horizon.
In addition to the roundtable discussion, participants were invited to share Recommended Resources from their work and studies to enrich further exploration of these topics.
Anton Korinek opened with a discussion of how advances in AI could bring us closer to artificial general intelligence and what that may or may not imply for labor markets. He shared preliminary results from a model he is working on to predict labor market outcomes from advanced automation in related scenarios with different starting assumptions about human capacity. In one scenario, humans are able to move to more complex jobs as a result of AI, in which case wages may increase indefinitely. In a scenario where there is a limit on the tasks that human brains are able to perform, in the near term, wages may increase due to productivity gains but then collapse as a result of the full substitution of human labor with machines.
Katya Klinova shared approaches to addressing the short-term disruption to the labor market anticipated as a result of AI implementation. Common approaches are upskilling and retraining, but she noted that these are reactive proposals that respond to disruption prompted by technology, as opposed to proactive efforts to design and implement more worker-friendly technology. To help various stakeholders be more intentional about AI deployment in the workplace, the Partnership on AI (PAI) has released Guidelines for AI and Shared Prosperity, which were built using insights from frontline workers and provide a high-level assessment and practical recommendations for AI developers, AI users, policymakers, labor organizations, and workers to mitigate risks of workplace AI.
Stephanie Bell presented more detailed findings from PAI’s research. She said that the problems from new technologies are largely predictable: increased job intensity, lower amounts of discretion, and higher rates of injury as a result of speed. However, PAI also found unanticipated problems that are becoming more prevalent. For instance, firms are automating tasks humans enjoy, and the use of AI is displacing the opportunity for workers to connect with other people and draw on uniquely human skills. Stephanie closed her comments by asking how we pursue expansion of the economy while leaving work humans enjoy.
Declining Job Quality: Lower labor demand may lead to lower job quality, and the implementation of artificial intelligence continues a downward trend in labor demand that’s been going on for decades as a result of other forms of automation.
Wage Decreases: Automation may economize labor in a way that has a significant negative impact on wages even though, in the immediate future, wages are likely to increase before the real impact is seen. Wage decline and job quality move in tandem and should be discussed together.
Poor Design: These technologies are not necessarily designed with the workers in mind. Workers often have no input in the design of new technologies even though they are most aware of friction points and what is likely to be successful for them. This can lead to worker frustration. Furthermore, the technology does not work well for everyone and leaves people with disabilities behind.
Worker Disempowerment: Workers are resigned to dealing with new technologies. Employers need to consider if they are really doing a service to employees by implementing new technology.
Loss of Data Rights: Data and privacy are as important as technological applications. Workers might not care that AI is being used, but do they know how data about them is being used? There is a lot of debate about regulation of who owns worker data.
Political Consequences: Displacement of workers could lead to social and geopolitical instability. Artificial intelligence has the potential to concentrate power in the state, threatening to destabilize democracies and pave the way for the rise of fascism.
Corporate Influence: The market is skewed to centralized power over AI, concentrating control in a few large corporations and disenfranchising small communities.
Economic Consequences: If AI leads to a reduction in labor, that will result in reduced tax revenues. Without tax revenues, how do we continue to fund public services?
Systemic Risk: AI will soon be embedded in everything, creating a risk that disruption within the industry could trigger the collapse of the entire economy similar to what happened in the financial crisis of 2008.
Quality of Life
AI has the potential to improve energy efficiency and clean energy production. It can be used to address gaps in health care services and improve access to health care. Artificial intelligence can also increase accessibility for people with disabilities.
AI has the potential to eliminate rote tasks, create better processes to bring workers into the loop of addressing known problems in work systems, and allow workers to be interdisciplinary instead of having to work solely within a specialty. Furthermore, AI can actually be used to measure job quality, which cannot be easily calculated using a single metric.
AI can increase productivity in expensive industries (e.g., healthcare, education), which had not seen recent boosts in productivity. If the implementation of new technologies leads to higher productivity, that results in greater societal wealth, making social services more affordable.
Reasons to be Hopeful
- Conversations like these need to be had—the earlier, the better—and they are happening quickly.
- We are moving towards realizing the AI Bill of Rights.
- Communities are vocalizing the need to establish standards for the implementation of artificial intelligence.
- Workers are demonstrating resilience in the face of technological change. We should not underestimate them.
Get to know the Speakers
Anton Korinek is a David M. Rubenstein Fellow at the Brookings Institution; a Professor at the University of Virginia, Department of Economics and Darden School of Business; a Research Associate at the National Bureau of Economic Research; a Research Fellow at the Center for Economic and Policy Research; and the Economics of AI Lead at the Centre for the Governance of AI. He received his Ph.D from Columbia University in 2007 after several years working in the IT and financial sectors. He has also worked at Johns Hopkins and at the University of Maryland and has been a visiting scholar at Harvard University, the World Bank, the IMF, the BIS, and numerous central banks. His research analyzes how to prepare for a world of transformative AI systems. He investigates the implications of advanced AI for economic growth, labor markets, inequality, and the future of our society.
Katya Klinova is the Head of AI, Labor, and the Economy at the Partnership on AI where she directs the strategy and execution of research programs to ensure AI supports an inclusive economic future. Prior to the Partnership on AI, Katya worked at the UN Executive Office of the Secretary-General on preparing the launch of the Secretary-General’s Strategy for New Technology. She also worked at Google on Chrome, Play, Developer Relations, and Search, where she was responsible for launching and driving the worldwide adoption of Google’s early AI-enabled services. Katya holds an MPA in International Development from Harvard University, a B.Sc, cum laude in Applied Mathematics and Computer Science from Rostov State University, and a Joint M.Sc in Networks and Data Science from University of Reading, Aristotle University of Thessaloniki, and Universidad Carlos III de Madrid, where she was a Mundus Scholar.
Stephanie Bell is a Senior Research Scientist at the Partnership on AI with the AI, Labor, and Economy team. She leads the design of corporate targets and commitments on Responsible AI, currently focusing on ensuring the technological gains from AI translate into equitable economic growth and high quality jobs. She is particularly interested in finding ways to include the perspectives of front-line workers in AI design and development, and exploring how AI can create better opportunities for low-wage workers. Prior to joining the Partnership on AI, Stephanie was an Engagement Manager at McKinsey & Company. She holds a DPhil in Politics and an MPhil in Development Studies from the University of Oxford, where she studied as a Rhodes Scholar. She also holds a B.A. in Anthropology from the University of Chicago, where she was selected as a Truman Scholar.
Thank you to all of the participants in the Working With AI roundtable for sharing both their time and wisdom. Their contributions were enlightening and invaluable and will shape conversations about the labor implications of AI to come.
Special thanks to our speakers, Anton Korinek, Katya Klinova, and Stephanie Bell, for their expert insight and for inspiring a lively discussion. Their willingness to share their knowledge was central to our shared understanding of the issues that need to be addressed.
Finally, thanks to my Aspen Digital team, Eleanor Tursman, B Cavello, and Carner Derron for their partnership in organizing and facilitating this conversation and sharing these insights with the world.
Elizabeth Miller is a Google Public Policy Fellow with the Aspen Digital program at the Aspen Institute, where she works on artificial intelligence-related policy issues. She is currently pursuing a master’s degree in Communication, Culture & Technology at Georgetown University, with a focus on safety implications of digital platforms. This fall, she will serve as a teaching assistant for a course on research methods. Elizabeth has a degree in communication, specializing in public relations, from the University of Maryland, College Park.