Flawed Logic in the Robots-Jobs Debate

By Peter Harris | November 15, 2017
Find me on:



At least in biopharma research, robots are less likely to replace jobs and more likely to increase human efficacy




Public discussion on the topic of robots taking jobs has to be at or near the zenith of the hype cycle. It seems like a new study on X% of workers losing their jobs to robots in Y industry is issued every day. Works like the Oxford University study in 2013, Martin Ford’s Rise of the Robots, and the Brynjolfsson/McAfee book The Second Machine Age have all helped to fuel the public perception that we stand on the edge of a technological precipice in which we may wake up one day to a world in which humans have nothing to do.

To some degree, the hype is rooted in real developments and disruptions coming as the result of very rapid developments in underlying technologies driving automation. Machine learning, sensing, collaborative robotics, cloud computing and general software advances are combining to open up a wide range of automation possibilities previously not possible. For more on our views on collaborative robotics specifically, see this article in Drug Discovery World.

With the above said, most analyses attempting to quantify job losses share a problematic presupposition: a static level of work needed/desired to be done. History has shown that this is never the case, as humanity constantly strives to improve standards of living through increased work, research, and innovation. As a result, we should instead expect that while many routine tasks performed by humans will end up being performed by robots, there is little historic precedent in assuming the aggregate amount of work will in fact decrease.

When the Luddites rebelled against the automation of textile and weaving machinery based on fears of losing jobs and livelihood, they shared the same logic of automation consuming required work on a static level of activity. They would certainly have been correct, if the result of the automation was simply to do the same level of work, but with machines. However, rather than having automated machines making each person one or two suits to own, today people have large closets full of many, many more items of clothing than would have ever been considered back in the early 1800s. The result of automation in this case was the expansion of production, not a simple consumption of fixed tasks by machines.

In biopharma research, it remains remarkable how much routine lab work is done on a daily basis by highly educated scientists and chemists. The cost of drug discovery has risen meaningfully over the last couple of decades, and the average number of new molecular entities approved per year has not kept pace. Is this because we have lost interest in improving our standard of living through novel therapies? Not at all.

As robots and automation continue to become critical fixtures in biopharma research, we should not expect a loss of jobs based on maintaining a fixed level of current activity. Rather, it is much more likely that robots and automation will be instrumental in driving substantially more experiments, of better quality, enabling scientists to do science, and more of it. This should in turn lead to better, more productive research output, not a reduction in research labor. Simply put, robots and automation are much more likely to increase the effectiveness of our research dollars rather than reducing research activities, and jobs. We look forward to what robots will bring in improving human health through making research and researchers more productive.


Topics: robots, jobs, lab automation, research, science