By 2026, 30 million places could be closed, according to an unpublished study that evaluated 2,602 occupations
The possibility of may be substituted for a robot or a computer program puts at risk 54% of the formal jobs in the country, shows an unpublished study done with Brazilian data.
By 2026, 30 million formal job openings would be closed if all companies decided to replace human workers with the technology already available – the number takes into account the trend of hiring for the most endangered occupations.
Considering the number of workers with a formal contract at the end of 2017 (according to the Ministry of Labor), about 25 million (57.37%) occupied positions with a very high probability (above 80%) or high (from 60% to 80%) of automation.
These categories include chemical engineers (96%), warehouse loaders (77%) and volleyball referees (71%), for example.
In order to calculate the probability of automation within ten years, 69 academics and machine learning professionals were consulted (artificial intelligence field in which computers discover solutions on their own after analyzing previous decisions).
From the evaluations of these experts, the researchers used techniques of analysis of the descriptions of the occupations, to associate the risks.
Machine learning boosts automation because it allows replacing not only repetitive and mechanical tasks such as elevator operator or typist (above 99%). Skin-based diagnosis of skin lesions is already done in less time and with more precision than those performed by humans.
The “tropicalization“ of the study is the first step, and the estimates still need to be refined and deepened, says UnB professor Pedro Henrique Melo Albuquerque.
And the probability of automation does not mean that, in practice, the human will lose a place. One of the bottlenecks, Albuquerque notes, is that machine learning feeds on a rich source of quality data rarely available to companies in developing countries.
Another limitation is economic. Virtual automation – such as implementing software to make laboratory test reports – require much less investment than buying robots for a jet launderer, for example.
Even when the technology exists and is accessible, replacement may not be feasible. This is the case of one of the most endangered occupations, the shorthand (99.55% probability of automation), whose professionals use codes to annotate up to 120 words per minute, then “translated“ into typed texts.
In theory, voice recognition programs would do that. In fact, they make so many mistakes that rechecking takes longer than typing, says Emília Naomi Todo Liem, manager of 13 stenographers at the São Paulo Legislative Assembly.
“In the future, if the number of hits grows, it can be a useful tool to avoid repetitive strain injuries. At the moment, it just makes you lose your temper, “she says.
Political pressures also influence. In 2018, the Post Office considered to extinguish a position that manually checks each package or letter and separates them according to the destination – the so-called operator of sorting and transshipment (75% risk, according to the study).
Automation, already used by competitors, was considered necessary to reduce costs and maintain competitiveness.
Striking workers pressed, however, and the position was maintained, with no deadline for termination. According to the company, it will happen when the 14,000 operators still in activity retire or “migrate to other compatible activities.“
There are cases where automation is successfully deployed, but jobs are maintained. Box operators, for example, are 77% likely to be replaced by machines. But despite already having self-service in 250 stores in Brazil, McDonald’s says there is no risk of layoffs.