Humans Still Cheaper Than AI in Vast Majority of Jobs, MIT Study Reveals

Recent Massachusetts Institute of Technology research has shed light on a crucial aspect of the ongoing debate regarding artificial intelligence (AI) and its role in the job market. Contrary to the rapid pace of AI development and adoption in various industries, the study reveals that AI is not yet a cost-effective replacement for human labor in most jobs. 

This finding is particularly relevant as it challenges the widespread apprehension about AI's potential to supplant human roles across diverse sectors. Focusing on specific applications where computer vision is essential, such as in educational and property appraisal roles, the researchers have provided a comprehensive analysis of the economic feasibility of AI in these areas. 

The study found that, in terms of dollar wages, only about 23% of workers could potentially be cost-effectively replaced by AI. For most tasks, the high costs involved in installing and operating AI-assisted systems make human labor the more economical choice.

The study, titled “Beyond AI Exposure,” conducted by MIT FutureTech along with The Productivity Institute and IBM’s Institute for Business Value, has provided a detailed evaluation of the viable tasks for automation, especially those involving computer vision. Researchers concluded that only 23% of wages paid for computer vision tasks would be cost-effective to automate with current AI technologies. This finding is significant in light of the accelerated adoption of AI across various industries, spurred by advancements in tools like OpenAI’s ChatGPT.

One of the key findings of the MIT study is the high cost associated with implementing and operating AI systems, particularly those that require advanced capabilities like computer vision. For instance, in industries like retail, transportation, and warehousing, where giants like Walmart Inc. and Amazon.com Inc. are prominent, the cost-benefit ratio of computer vision might be more favorable. However, the economic feasibility of replacing human labor with AI in many other sectors is still questionable.

While the study suggests that AI will not imminently replace human jobs on a large scale, it does not rule out the possibility of this occurring in the future. The paper indicates that AI could potentially overtake human roles if the costs of deployment decrease or if AI is deployed through AI-as-a-service platforms, which offer greater scalability than individual firm deployments. However, this scenario is predicated on significant advancements in AI technology and reduced associated costs.

The research includes a practical example of a small bakery, where the cost of automating tasks like quality control of ingredients with computer vision systems far exceeds the labor costs saved. This highlights the economic unfeasibility of AI in small-scale operations and tasks that comprise only a minor portion of the overall job responsibilities.

The MIT study offers a nuanced view of the AI-labor dynamic, suggesting that the shift toward automation will be more gradual rather than abrupt. The current economic model shows that the job displacement due to AI, even within the realm of vision tasks, will likely be smaller than the existing job churn in the market. This underlines the importance of considering both the technical feasibility and the economic practicality of AI systems in any discussion about automation and job displacement.