Credit: Massachusetts Institute of Technology, Midjourney
Like many of us, you might find yourself nodding to the familiar digital doomsday refrain that resonates in offices and coffee shops alike: AI will take my job!
Is this looming threat justified or simply a manifestation of our shared anxiety over constant technological advancement? A new study from MIT CSAIL, MIT Sloan, the Productivity Institute and IBM’s Institute for Business Value should challenge our long-held beliefs.
Their research critically examines the economic feasibility of using AI to automate tasks in the workplace, with a particular focus on computer vision.
Their results show that currently only about 23% of salaries paid for tasks involving vision are economically viable for AI automation. In other words, it only makes economic sense to replace human labor with AI in about a quarter of jobs where vision is a key part of the job.
“This indicates a more gradual integration of AI across various sectors, contrasting with the rapid displacement of jobs due to AI, often hypothetically,” says Neil Thompson, a senior fellow at MIT CSAIL and the AI Initiative. Digital Economy. “We focused on the area of computer vision, an area where cost modeling has seen significant progress.”
The study departs from the conventional general approach to the potential impact of AI. Instead, it offers a careful examination of the feasibility of AI in automating specific tasks. What sets this research apart is its tripartite analytical model. The framework not only assesses the technical performance requirements of AI systems, but also delves deeper into the characteristics of an AI system capable of achieving this performance, as well as the economic choice of building and deploying such a system.
Many years of experience in the computer vision field provide abundant data for evaluating performance and economic viability. On the other hand, data relating to new major language models are still developing. Fortunately, experience with computer vision models provides insight into what the future might hold with the spread and adoption of language models. According to the researchers, development, deployment and operating costs could decrease and the technology industry could transform to provide AI solutions as a service, eliminating the need for large investments.
Researchers examined the implications of potential cost reductions in AI systems and how these changes could influence the pace of automation. For example, if the costs of implementing AI in the workplace decreased significantly, this could accelerate the pace of AI adoption across various industries, potentially leading to more rapid changes in the job market. work. Conversely, if computing needs increase, data becomes harder to find, and skilled workers become scarce, higher costs could slow this transition, giving workers and industries more time to adapt. adapt.
Another critical aspect: AI-as-Service platforms. The scientists showed how scalability and broader application could potentially change the landscape of task automation, shifting the focus from individual, enterprise-level deployment to a broader, service-based approach. “The implications of this shift are profound: it could democratize access to AI technologies, allowing small businesses and organizations to benefit from AI without the need for significant internal resources. Additionally, this could lead to the emergence of new business models focused on AI services,” says Thompson.
“When the semiconductor industry created an entirely new business model 20 years ago with the separation of design and manufacturing with outsourcing of production, fabless semiconductor companies became the standard,” says Martin Fleming, former IBM chief economist and director of analytics and now an IBM Fellow. The UK-based Productivity Institute says, “In the coming years, its eventual software, cloud services and consulting companies will create a new business model with a class of companies specializing in AI as a service at scale . »
The implications of the study extend beyond immediate economic considerations and touch on broader societal impacts such as workforce retraining and policy development. It paves the way for further research into the scalability, cost-effectiveness and ability of AI to create new job categories. As some jobs are automated, for example, there will be a growing need for roles focused on managing, maintaining and improving AI systems, as well as roles in areas where human skills are irreplaceable by AI.
Furthermore, to the extent that AI cost reductions, new AI services, or both, succeed in contributing to better macroeconomic productivity growth, employment growth, and income growth. will accelerate and the standard of living will improve. “Broad economic benefits will only be realized when a fundamental transformation occurs in the way business is done and the way workers work,” says Fleming.
Comparison of AI exposure and firm-level economic attractiveness for computer vision. Credit:
New economic models are beginning to emerge. For example, small jewelers benefit from a diamond grading tool built by NavTech in which an image provided by a jeweler is graded to instantly establish quality without the availability of an experienced jeweler.
For autonomous vehicles, Nvidia has built a platform using high-performance computing, imaging and AI enabling continuous improvement and deployment via over-the-air updates. Individual automakers no longer need to create duplicate features such as stereo camera and route recognition technologies.
“As AI continues to advance and reshape industries, we hope the results of this study will serve as a crucial benchmark, guiding future explorations and policy development at the ever-evolving intersection of technology, economy and labor market to help us navigate the future. the challenges and opportunities presented by the continued integration of AI into the workplace,” says Thompson.
“Much has been written about the future impact of AI on the labor market, primarily using exposure metrics. However, these estimates often rely on the assumption that whether a job can be automated , it will be,” explains Antonin Bergeaud, associate professor. in economics at HEC Paris.
“The research by Svanberg and his co-authors takes a new perspective by meticulously estimating the costs of implementing these technologies, from installation to maintenance. It reveals that even an AI system that is “only” as good as a human would often be prohibitively expensive. adopt, relative to current labor costs in the United States
“The conclusion is striking: a much smaller share of the labor market is at risk of automation than direct exposure-based estimates suggest. This important result calls for a more systematic assessment of the feasibility of adopting a new technology for an industry, which directly relates to the new Solow paradox, in which companies may fail to adopt an overperforming technology if the barriers are too high.
More information:
Beyond Exposure to AI: What Tasks Are Profitable to Automate with Computer Vision? futuretech-site.s3.us-east-2.a … yonde_AI_Exposure.pdf
Provided by the Massachusetts Institute of Technology
Quote: Rethinking the impact of AI: a study reveals the economic limits of task automation (January 23, 2024) retrieved January 23, 2024 from
This document is subject to copyright. Apart from fair use for private study or research purposes, no part may be reproduced without written permission. The content is provided for information only.