AI costs more than human labour, according to MIT

Social Media Plug-ins:

The Massachusetts Institute of Technology (MIT) has revealed that it is still cheaper to use humans for certain jobs than artificial intelligence (AI).

In a study addressing fears about AI replacing humans in various industries, MIT established that using AI to replace humans is only profitable in a few industries.

This is coming amid concerns that AI will replace many jobs currently handled by humans. The report suggests that AI cannot replace most jobs in cost-effective ways at present.

Already, the International Monetary Fund (IMF) has said that almost 40 percent of jobs across the globe will be influenced by AI, projecting that advanced economies are expected to experience a higher impact compared to emerging markets and low-income nations.

In a blog post, IMF chief Kristalina Georgieva called for governments to establish social safety nets and offer retraining programmes to counter the impact of AI.

But MIT explained that the cost-benefit ratio of computer vision is most favourable in segments like retail, transportation, and warehousing—areas where Walmart Inc. and Inc. are prominent—and in the healthcare context.

It said computer vision is a field of AI that enables machines to derive meaningful information from digital images and other visual inputs, with its most ubiquitous applications showing up in object detection systems for autonomous driving or in helping categorise photos on smartphones.

Giving further insights, MIT researchers said they found that AI could effectively supplant only 23 percent of workers, measured in terms of dollar wages.

In other cases, because AI-assisted visual recognition is expensive to install and operate, humans did the job more economically.

“Machines will steal our jobs” is a sentiment frequently expressed during times of rapid technological change. Such anxiety has re-emerged with the creation of large language models, said the researchers from MIT’s Computer Science and Artificial Intelligence Laboratory in the 45-page paper titled “Beyond AI Exposure.”

“We find that only 23 percent of worker compensation ‘exposed’ to AI computer vision would be cost-effective for firms to automate because of the large upfront costs of AI systems.”

The study was funded by the MIT-IBM Watson AI Lab and used online surveys to collect data on about 1,000 visually assisted tasks across 800 occupations.

Only three percent of such tasks can be automated cost-effectively today, but that could rise to 40 percent by 2030 if data costs fall and accuracy improves, the researchers said.

Commenting on the study, the Director of the FutureTech Research Project at the MIT Computer Science and Artificial Intelligence Lab, Neil Thompson, said: “Our study examines the usage of computer vision across the economy, examining its applicability to each occupation across nearly every industry and sector. We show that there will be more automation in retail and healthcare and less in areas like construction, mining, or real estate.”