Many economists and technologists believe the world is on the brink of a new industrial revolution, in which advances in the field of artificial intelligence will obsolete human labor at an unforgiving pace. Two Oxford researchers recently analyzed the skills required for more than 700 different occupations to determine how many of them would be susceptible to automation in the near future, and the news was not good: They concluded that machines are likely to take over 47 percent of today’s jobs within a few decades.

This is a dire prediction, but one whose consequences will not fall upon society evenly. A close look at the data reveals a surprising pattern: The jobs performed primarily by women are relatively safe, while those typically performed by men are at risk.

It should come as no surprise that despite progress on equality in the labor force, many common professions exhibit a high degree of gender bias. For instance, of the 3 million truck drivers in the U.S., more than 95 percent are men; of the nearly 3 million secretaries and administrative assistants, more than 95 percent are women. Autonomous vehicles are a not-too-distant possibility, and when they arrive, those drivers’ jobs will evaporate; office-support workers suffer no such imminent threat.

This pattern holds for many of the most gender-biased occupations. Men hold 97 percent of the 2.5 million U.S. construction and carpentry jobs. The Oxford study estimates that these male workers stand more than a 70 percent chance of being replaced by robotic workers. By contrast, women hold 93 percent of the registered nurse positions. Their risk of obsolescence is vanishingly small: .009 percent.

What is causing this pattern? The skills exhibited by the coming wave of intelligent machines are better suited to occupations currently dominated by men. Many of the jobs held by men involve perception and manipulation, often in conjunction with physical exertion, such as swinging a hammer or trimming trees. The latest mobile robots combine advanced-sensory systems with dexterous manipulators to successfully perform these sorts of tasks.

Other, more cerebral male-dominated professions aren’t secure either. Many occupations that might appear to require experience and judgment—such as commodity traders—are being outdone by increasingly sophisticated machine-learning programs capable of quickly teasing subtle patterns out of large volumes of data.

By contrast, women typically work in more chaotic, unstructured environments, where the ability to read people’s emotions and intentions are critical to success. If your job involves distracting a patient while delivering an injection, guessing whether a crying baby wants a bottle or a diaper change, or expressing sympathy to calm an irate customer, you needn’t worry that a robot will take your job, at least for the foreseeable future.

So what will the new machines be good at? For starters, they will be well-suited to tasks that are easily specified and offer objective criteria for success. These features permit an engineer to codify requirements in a programmatic form and measure the results. It’s easy to understand what a robotic housepainter is supposed to accomplish and to see if the job has been done correctly; it’s harder to assess whether a dementia patient might be more comfortable with a warmer blanket. Computers also excel at tasks that benefit from consistency, attention, and objectivity, such as washing windows, managing the flow of air traffic, or assigning taxi drivers to trip requests.

Another characteristic affecting a job’s security is the breadth of skills it requires. Computers aren’t usually designed to replace workers; they typically automate specific tasks, making a given worker more productive. But when an automated system can match the entire range of that worker’s talents, his or her services are no longer needed. So the broader and more varied your duties, the harder it will be to replace you.

In short, today’s typical women’s work is what will predominate in future. On a mass scale, this pattern may result in an involuntary shift in the division of labor, with husbands tending to household duties after dropping their wives off at the office. Superficially, that may sound cheery, but the reality will be much grimmer, as families struggle to make ends meet on one income, and men struggle with the emotional upheaval of no longer having a place in the world of work.