The MIT study reveals a crucial insight into the cost-effectiveness of AI in job automation. It found that only 23% of worker compensation exposed to AI computer vision would justify the large upfront costs of implementing AI systems. In other words, the majority of tasks that AI could potentially automate remain economically unattractive for companies.

Conversely, 77% of vision-related tasks are deemed uneconomical to automate when AI systems are confined to firm-level usage. This stark contrast underscores the pivotal role that cost-effectiveness plays in the widespread adoption of AI technology.

Even if an AI computer system were priced as low as $1,000, there remain tasks in the job market that are simply not economically attractive to replace. Low-wage occupations and work within small firms fall into this category, illustrating the complexity of job displacement in the AI era.