Artificial intelligence is often perceived primarily as an automation tool, but MIT economics professor David Autor argues that it should be viewed as a collaboration tool that enhances employee skills rather than replacing them. “For airline pilots, we really want them to be able to fly a plane manually. We don’t want them to be fully dependent on autopilot. So it matters whether their skills decline,” Autor said during a recent episode of the “MIT CSAIL Alliances” podcast.
The podcast, hosted by Kara Miller, delves into the impact of AI and computer science on work, education, and daily life. In this episode, Autor and MIT Sloan principal research scientist Neil Thompson discussed AI’s implications for jobs, productivity, and the future workforce. Their conversation yielded several insights, particularly the mixed messages surrounding AI’s effect on productivity.
Thompson cited a study scheduled for release in 2025 that examined experienced open-source developers writing an update to a software library. Although those who utilized generative AI completed their coding tasks more rapidly, they overall took 19% longer to finish the entire task compared to a control group. The additional time was spent on writing prompts, verifying outputs, and waiting for the AI model’s responses. Interestingly, the developers believed the AI tool had increased their speed by at least 20%. “This is not that I think that everyone is going to be made less productive by AI,” Thompson noted, pointing out the potential friction points in the automation process.
Automation can have varying impacts on labor value. In a paper authored by Autor and Thompson, they explored how automation accentuates the value of more advanced skills in the job market. For example, tools like spellcheck and autocorrect have elevated the worth of proofreaders, resulting in wage increases for those who remained employed in that sector. Conversely, when automation encroaches on expert tasks, wages can decline, as demonstrated by the impact of GPS technology on taxi drivers. “Someone can now walk in off the street and drive a taxicab pretty well,” Thompson said, emphasizing the need to assess whether automation enhances efficiency for skilled tasks or commoditizes them.
Autor asserted that collaboration is the ultimate goal of automation, as it introduces capabilities beyond human capacity. He referenced CheXbert, an AI model designed to analyze and label radiology reports, which performs comparably to two-thirds of practicing radiologists when making diagnoses based solely on X-rays. However, radiologists often performed worse when employing CheXbert, especially when the AI was uncertain about its conclusions. “It’s not a limitation of AI. It’s a challenge of designing it in a way that collaborates effectively with human capacities,” Autor explained.
Thompson further emphasized the financial implications of AI accuracy, noting that the cost of achieving higher accuracy levels can escalate drastically. “Many businesses I think are getting into this world where they say, ‘I want to fully automate,’ and then they realize that it’s just too expensive to do that,” he said. Instead, companies might consider retaining human oversight to review AI outputs during critical decision-making processes.
Towards the conclusion of their discussion, Autor and Thompson highlighted compelling use cases for AI, particularly in enhancing communication and operational efficiency. For instance, AI voice calls are being employed by courier services in China to facilitate better interactions between deaf or hard-of-hearing drivers and customers, resulting in fewer complaints and improved productivity and wages for drivers. Similarly, Google has utilized deep learning to optimize cooling in its data centers, significantly reducing energy consumption by adapting to various environmental factors.
Both Autor and Thompson bring extensive expertise to this dialogue; Thompson serves as a principal research scientist at MIT’s Computer Science and Artificial Intelligence Laboratory and is also the director of the MIT FutureTech lab. Autor is a co-director of the MIT Stone Center on Inequality and Shaping the Future of Work, focusing on the labor-market impacts of technological change. As AI continues to evolve, understanding its collaborative potential and nuanced effects on productivity and labor markets will be critical in shaping the future of work.
See also
AI Study Reveals US, China, and EU Diverge on Future Development Pathways
Fujitsu Launches AI-Driven Platform, Achieving 100-Fold Productivity Increase in Software Development
Georgia Tech Researchers Unveil ZEN Framework for 100% AI Model Attribution Accuracy
Brands May Gain Consumer Trust from AI Content, Study Finds Positive Reactions in 41%
ASML Unveils Lithography Breakthrough Boosting Chip Production by 50% for AI Semiconductors
