Industrialization disrupts the labour of artisans and craftsmen, just as the age of information technology and automation shifted tasks in the labour force and led to occupational obsolescence.
How will the rapid advances in artificial intelligence (AI) and related technologies affect workers? Which roles will be replaced? How does AI improve and enhance functionality? What valuable new collaborative skills does AI enable?
To explore these questions, Northwestern University’s Human Computer Interaction + Design Center (HCI + D) hosted a virtual panel with David Auto, a professor of economics at the Massachusetts Institute of Technology, and Eric Horvitz, Chief Science Officer at Microsoft.
Both Autor and Horvitz contribute authors to the national scholars of science’s “Artificial Intelligence and the Future of Work.” This assesses the current and future economic, productivity and labor impacts of AI across the sector.
“AI has spurred an unprecedented wave of experiments, but the future is inevitable,” said Liz Gerber, a professor of mechanical engineering at the McCormick School of Engineering and a coder at HCI+D, professor of communication studies at the School of Communications. “We have a role and can create new forms of valuable and productive work. People are making decisions and it is our duty to be considerate about them.”
Complement human expertise with a new mode of collaboration
Among the 11 key findings in the “Artificial Intelligence and the Future of Work” report, AI is classified as a general-purpose technology (similar to electricity and steam engines) with a wide range of applications and is likely a key driver of long-term economic growth.
Autor and Horvitz highlight the degree of uncertainty regarding the economic trajectory of AI, but they agree to restructure the labour market, albeit unevenly.
AI is committed to transforming many cognitive tasks across the industry and creating new efficiencies. AI can also automate or move specific tasks, but it could also create new roles and forms of work, particularly by complementing human expertise with AI capabilities in decision-making and complex problem solving. As AI changes roles as it handles repetitive, data-heavy tasks, workers can focus more on creative, interpersonal and strategic responsibilities.
“The thing that I think is most valuable is to supplement human judgment, especially when there are high interests, but there’s no clear answer,” Autor says. “Whether it’s patient diagnosis, kitchen remodeling, some kind of skill, repair, or software architecture, what’s really good about AI is that it provides guidance and guardrails on how to do that.”
“We barely damaged the surface of human AI’s collaborative possibilities,” says Horvitz, who launched Microsoft’s AI Anthology, a collection of essays on the future of AI. “It’s a very wonderful field for today’s research. It gives you a deeper look at new forms of collaboration that support the superiority of human institutions and contribution.”
In addition to increasingly enhancing human expertise more effectively, Autor and Horvitz propose that AI can help engage in traditional specializations such as coding, legal research, and healthcare diagnosis.
From a new type of work perspective, Autor and Horvitz warn against the failure of the imagination. Most technologies are valuable because they can do things we couldn’t do before, not because they allow us to do things we couldn’t do better or faster, but because they can enable us to do things we couldn’t do before.
“The Apollo Guidance computer was a small part of computer hardware that had less capacity than the typical modern washing machine,” Autor said. “Even so, it welcomed people on the moon.”
Target domains with high impact
Autor and Horvitz discussed the potential for transformation of sector-specific AI applications in education, healthcare and frontier science.
Horvitz is excited by the advances in biological sciences made possible by AI, from the advances in biological sciences made possible by AI, from drug discovery and reuse to predicting protein structure and function, to the synthesis of new reagents and molecules that play a role in the fight against chronic autoimmune diseases and cancer.
“You build a model that can expand the viable space for candidates that could be much larger than we’ve thought in the past,” explained Horvitz. “And we apply the relevant AI technology to adopt that expanded set of possibilities and quickly and efficiently hone the candidates most likely to have the right traits.”
Horvitz recently designed the AI virtual cell in a project with a team that includes Shana Kelley from Northwestern Engineering. It is a multi-scale, multimodal, large network-based model that can simulate the behavior of Silico molecules, cells, and tissues across diverse states.
Looking ahead
Autor and Horvitz advise on the on-the-season laissez-faire approach to AI, highlighting the need for proactive intervention through policy, governance, incentive structures, and training to proactively direct the future of AI in the labor market.
“We need to closely monitor the impact of AI on employment in the economy and creatively pursue policies that shape technologies that promote shared concepts of prosperity rather than increasing economic disparities,” says Horvitz.
Autor cited his friend, the philosopher Joshua Cohen. It’s a design exercise.
“Our job isn’t just to look ahead and say, ‘What’s going to happen?’,” Auto said. “We all intend to live with the outcomes, so we should recognize that we are people making decisions.”
Supported by Northwestern Engineering and Northwestern Communications Schools, HCI+D brings together researchers and practitioners from across the university to research, design and develop the future of human-computer interactions at home, at work and play. In addition to Gerber, the center is co-led by Darren Gergle, a BAO family professor of interaction with human computers in communication schools, and Bryan Pardo, a professor of computer science at Northwestern Engineering.
