Is a robot coming for your job? Now you can find out via an algorithm, appropriately. A handy calculator from NPR, based on research from Oxford University, will tell you how likely it is that your job will be automated in the next couple of decades.
At highest risk are jobs like telemarketing, insurance underwriting, and accounting. But dental technicians, legal assistants, and models are also near the top of the list, according to the algorithm. The study analyzed 702 different jobs, looking at characteristics like creativity and dexterity, and improvements in automation that are likely to happen in the next couple of decades.
Technology is changing quickly: As recently as 2004, some experts were saying that driving was too complex a task to be automated. Six years later, Google announced its first prototypes of self-driving cars.
Even restaurant servers have a 94% chance of being computerized, though the researchers originally expected that it was an example of a non-automatable job. “We were obviously thinking about the fact that waiters and waitresses require making small talk with customers at tables, presenting a smiling face in a way that robot isn’t able to do,” says Michael Osborne, one of the researchers. “So we told the algorithm that waiters were non-automatable, and yet, despite that, the algorithm came back and told us that the probability was .94.”
Now, two years after the research was first published, some waiters are already being replaced: Some restaurant chains, like Olive Garden, are now using tablets to take orders. The algorithm flagged the job because it doesn’t take a lot of originality–so it’s a prime candidate for robotic replacement.
The researchers identified several factors that increase the chances a job can be computerized. The more you’re required to personally help other people, the less likely you’re going to be replaced by a robot (though things like robotic nurses already exist). If your job requires negotiation, or a high degree of creativity, there’s also less risk.
Still, even jobs that might not be immediately threatened will likely use robotic assistance, especially to process big data. At hospitals, some oncologists are already starting to use computers to comb through millions of patient records and pages of medical journal text to find patterns and recommend specific treatments for individual patients. At law firms, algorithms are starting to scan legal briefs in place of paralegals or contract lawyers.
While the Oxford study doesn’t claim to know exactly when specific shifts will happen, it suggests that automation will happen in a few major waves. First to go: jobs in production, administration in offices, and transportation and logistics. Many service jobs, certain sales jobs (like telemarketing and cashiers, which are already starting to disappear), and construction jobs will also be automated soon.
The second wave of automation will depend on how quickly robots can learn creative and social intelligence–something that’s likely to take more time, but will probably happen eventually. “I think we can almost guarantee that technology will continue to progress and will ultimately render almost all the jobs that humans do today automatable,” says Osborne.
For now, those with the highest-skill, highest-paid jobs are probably safe, and low-skill workers are not. “Inequality is probably the foremost challenge,” says Osborne. “It’s not going to be a problem of there not being enough wealth. We’re fairly confident that all of these technologies will continue to generate vast amounts of wealth–we’ll be generating a cornucopia of increasingly cheap and wonderful goods that will be able to be produced for next to zero marginal cost. But those benefits we’ll see as consumers might not necessarily be realized by workers.”
As jobs begin to disappear, those with the least education will have the challenge of figuring out what to do next. “We’ve got this fear that the burden of this automation might rest most heavily on the shoulders of those who are least skilled, and hence perhaps less well equipped to move into whatever new occupations are generated,” he says.
While some suggest that a universal basic income might be the answer, the researchers think it’s too early to tell. “I haven’t really seen a suitable solution to this issue yet,” says Osborne. “In the case of universal basic income, there are a range of political challenges to getting such a thing instituted. We’ll have to see.”
An algoirthm is probably not ready to supply answers either. You can use NPR’s calculator to check out the risk of your job being automated here.