Human workers still more cost-effective than AI in most jobs - MIT study
A study from the Massachusetts Institute of Technology (MIT) found that only 3% of visually-assisted tasks can be automated cost-effectively currently, but that could rise to 40% by 2030 if data costs fall and accuracy improves, according to Bloomberg and CNN.
The study focused primarily on jobs using computer vision, a part of AI that helps machines understand important details from pictures and other visual elements. It is commonly used in autopilots, or sorting pictures on phones. MIT's Computer Science and Artificial Intelligence Lab used online surveys to gather data on nearly 1,000 visually-assisted tasks across 800 occupations.
“We find that only 23% 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 researchers say in the study.
Most jobs currently considered at risk of being taken over by AI are not economically profitable, as the study shows. At the moment, it makes more sense for companies to keep humans doing these jobs.
The study shows that, despite all the talk about robots taking over jobs, it's not going to happen right away. Neil Thompson, one of the researchers, explained that even though AI has the potential to replace some tasks, it's not economically smart for companies to do it immediately.
What jobs are at risk
The study analyzed different jobs and found that AI is most helpful in stores, especially for big companies like Walmart and Amazon, making a retail worker who checks inventory replaceable by a machine with computer vision. AI can also work well in healthcare, according to MIT's paper.
However, it is still more cost-effective for a company to have a human do these tasks, the study says. The authors suggest that if we push AI more, especially through subscription services, it could become more useful in other ways too. Less automation is expected in construction, mining, or real estate.
“In many cases, humans are the more cost-effective way, and a more economically attractive way, to do work right now,” Neil Thompson, one of the study’s authors. told CNN.
Thompson pointed out that the slow adoption of AI in the workforce is similar to how other technologies took time to become widespread. Just like the shift from agricultural to manufacturing economies happened gradually, the impact of AI tools on jobs will likely be a gradual process.
This research suggests that there is more time for governments and businesses to plan for the changes AI might bring to the job market. Policymakers can use this data to make plans for retraining workers or establishing safety nets to protect people from job disruptions caused by AI.
AI impact on the workforce
After the Microsoft-backed OpenAI company introduced ChatGPT, an AI-powered chatbot, a wave of similar developments arose, including Google Bard, Google Gemini, and Samsung's AI Gauss.
A global trend in implementing AI into business has already started, making both employees' work and customer service more efficient. For example, Walmart implements AI for its new shopping features to make shopping in stores or online feel more advanced.
The International Monetary Fund predicted that almost 40% of jobs worldwide could be replaced by artificial intelligence. In more developed countries, artificial intelligence could affect 60% of jobs, while in emerging and low-income countries, AI is expected to affect 40% and 26% of jobs, respectively.