I teach a course about the future of healthcare to medical, public health, and allied students. In one lecture I ask students to design the future of care. They come up with their own ideas. We talk them through those and design the process of care in real time. Doing this I learn a lot each semester about how students think about the future. I had an older student with a previous degree in economics. He had decided to become a doctor at the age of 30. When I spoke about what jobs robots and algorithms might take in the future, he raised his hand and asked whether robots would take over our jobs in healthcare–with a very worried look on his face.

Surgical robots become increasingly precise each day. Man–size robots can lift and move patients and transport them throughout the hospital. I held a PARO therapeutic robot in my arms. It was cute and calmed me. At a conference I once watched how a diminutive robot made an entire audience dance with it. It only takes the Xenex robot 10 minutes to disinfect a patient room with UV light. A robot called Tug works at hospitals in the San Francisco Bay Area. It delivers food and medicine. It picks up waste and laundry. It navigates the halls without crashing into people.

Paro robot
Paro robot

The above student asked about robots, but I think he was really asking about automation. Automation includes robotic devices, robots that look like a human, and algorithms. Silicon Valley investor Vinod Khosla once said something that resonated within the medical community for a long time. He said that technology would replace 80% of doctors because machines, driven by big data and computational power, would not only be cheaper but more accurate and objective than the average doctor. He added that we eventually wouldn’t need doctors at all.

In 2015 the information technology research firm Gartner predicted that one–third of existing jobs will be replaced by software, robots, and smart machines by 2025. Blue collar as well as white collar workers such as financial and sports reporters, marketers, surgeons, and financial analysts were in danger of being replaced. As Martin Ford outlines in Rise of the Robots, healthcare represented less than 6% in the US economy in 1960. Its share had tripled by 2013. The real issue is not utilizing too many robots but too few. Typically robots are expensive but reduce costs. Medicine and healthcare won’t be able to and should not try to avoid this.

If we look at the history of automation the first wave of machines in the 19th century was better at assembling things than people were. The second wave machines were better at organizing things. Today data analytics, cognitive computers, and self–driving cars suggest that they are better at pattern–recognition.

But both the simplest tasks and the most complicated ones require people. By simplest I mean that there is a greater chance a robot can play chess than go upstairs. By complicated I mean that regarding jobs such as managers, healthcare workers, and others related to education or media; humans are still superior at working with, and caring for others humans. Although, making a diagnosis is cheaper with cognitive computers than doing that alone as physicians.

But whether a robot can make an ethical decision is a huge question. An interesting experiment raised this question. In it a small robot was programmed not to let other robots called human proxies, which represented real people, get into the danger zone on a table game. When only one human proxy approached the danger zone, the robot could successfully thwart it. But when two proxies appeared the robot became confused, and in 14 out of 33 trials it wasted so much time trying to decide that both human proxies fell into the hole. Robots cannot make yet the ethical decisions that characterize experienced physicians.

A Robot companion for the elderly
A Robot companion for the elderly

Automation will make the world better and create opportunities for people clever enough to seize them. But healthcare will change. Tasks and procedures that can be automated should be, and will be. Algorithms will make diagnoses based on quantifiable data better than how humans do it now alone. It is easy to automate the fabrication of equipment or the transportation of patients. The challenge comes when empathy and interpersonal interaction comes into play. Robots won’t approach this level of sophistication for a long time.

To answer my initial question: many jobs will be taken over by robots and automation in the coming years. If people whose jobs are replaced cannot acquire new skills or improve their existing ones, they will no longer have a job. Given this possibility we must constantly question what our best individual skills are and what we can do to improve them. Let’s make sure to attend to those skills that make us irreplaceable.

Read more about other 39 exciting questions in My Health: Upgraded.