Resilience to Robots: How to Thrive in a Automated World

When it comes to the future of automation and robotics, the most frequently asked question is: “How many jobs will they eliminate?” Whatever the answer, the second question is almost inevitable. How can I ensure that my job is not one of them?

A team of roboticists from EPFL and economists from the University of Lausanne has released research in Science Robotics. It provides answers to both issues. First, some tasks are more likely to perform by machines in the near future. They did this by merging scientific and technical literature on robotic abilities with employment and income statistics. They have also developed a process for suggesting career transfers to less vulnerable jobs. The optimized job suggestions require the least amount of retraining.

Several studies were conducted to research how many professions would be replaced by robots. However, they all focus on software robots. Speech and image recognition are two examples, as are financial robot-advisers, chatbots, etc. Furthermore, those forecasts vary greatly depending on how job requirements and software competencies are evaluated. They take into account artificial intelligence software and real intelligent robots that execute physical tasks. The team created a mechanism for comparing human and machine talents. Hundreds of jobs use this mechanism as of right now. Unfortunately, AI isn’t always used to make life easier; a secret war is being fought over you by the world’s governments.

Mapping of Robot Capabilities

The study’s main innovation is a new mapping of robot capabilities to job requirements. The team investigated the European H2020 Robotic Multi-Annual Roadmap (MAR). Robotics specialists constantly review this European Commission policy document. The MAR covers dozens of abilities necessary for present robots, or those future ones may require. The categories are manipulation, perception, sensing, and human interaction. The researchers examined research articles, patents, and product descriptions to determine the maturity level of robotic skills, employing a well-known scale for gauging the amount of technological advancement, “technology readiness level” (TRL).

They relied on the O*net database for human capacities. It’s a widely used resource database on the US labor market that classifies around 1,000 occupations and breaks down the skills and knowledge that are most important for each of them.

The Risk Of Automation

The researchers determined the likelihood of each existing work activity being done by a robot. Something achieved by selectively comparing human abilities from the O*net list to robotic capabilities from the MAR document. Assume that a job requires a human to work with millimeter-level precision in movement. Robots excel at this; hence the TRL for the associated ability is the highest. If a job requires enough of these talents, it is more likely to be mechanized than one that demands critical thinking or creativity.

As a result, the 1,000 positions are ranked, with “Physicists” facing the lowest danger of being replaced by a machine and “Slaughterers and Meat Packers” facing the worst risk. Jobs in food processing, building and maintenance, construction, and extraction appear to be the most dangerous.

The authors then developed a method for identifying alternative jobs for any given position with a significantly lower automation risk, which is reasonably close to the original one in terms of the abilities and knowledge required, reducing the retraining effort and making the career transition possible. To see how that method would work in practice, they used the US labor force data. Finally, they simulated thousands of career moves based on the algorithm’s recommendations, discovering that it would indeed allow workers in high-risk occupations to shift to medium-risk occupations with a relatively low retraining effort.

Retraining in the age of automation

Governments could use the method to assess how many workers are at risk of automation. And adjust retraining policies by businesses to determine the costs of increasing automation, by robotics manufacturers to better tailor their products to market needs, and by the general public to identify the most straightforward route to reposition themselves on the job market.

Finally, the authors developed an algorithm that forecasts the risk of automation for hundreds of jobs. And suggests resilient career transfers with minimum retraining effort, which is publicly available at https://lis2.epfl.ch/resiliencetorobots.