The latest from the Bureau of Frightening Statistics is that within 20 years, almost half of all U.S. jobs will be automated. Robots and artificial intelligence (AI) routines are now able to tackle many routine tasks. The technology continues to improve, tackling more and more routine tasks and displacing the workers employed doing those tasks. I don’t fancy myself a Luddite (a person opposed to increased industrialization or new technology), and I have embraced the benefits of technology at work and in my personal life. My thermostat adjusts itself to save energy. The microwave knows how to warm foods best, but I still have to tell it what the food is. My house could be vacuumed by a robot, though it currently isn’t. I time shift TV programming to fit my schedule. My new car actually watches the road, adjusting speed, watching my blind spots and telling me if I begin to venture out of my lane.
We are, it is said, on the verge of an inflection point. Self-driving cars are on the road today, harbingers of a driverless world. I attended an ACS strategy session today and experienced Washington, D.C., traffic. A world of self-driving cars is certainly appealing. The time I just wasted in traffic on I-95 would have been put to better use. However, wide spread use of self-driving cars would also mean those professional drivers that were trapped in traffic with me would be out of work. There are 1.7 million truck drivers that stand to lose their jobs when self-driving trucks become the norm. Driving is pretty difficult, and I know from personal experience that driving an 18-wheeler is even more challenging. If drivers can be replaced, what other careers are in jeopardy as a result of new technologies?
It is a long list. A list that is beginning to be revealed as telemarketers, fast food employees, insurance underwriters, warehouse workers, and tax preparers join buggy-whip makers, record store employees, library workers, and photo technicians--occupations made obsolete by the march of technology. The list of obsoleted occupations is sobering, and the future promises that robots and AI will be taking on more challenging, higher skilled tasks. Robots perform surgery as computers diagnose disease. Doctors, it would seem, may be at risk for replacement. Physicians, chess champions, and poker players all could be replaced by computers.
Robotics long ago invaded the chemistry lab. Recent advances have me pondering whether chemistry’s Luddite moment is ahead of us or behind us? Has chemistry embraced robotics and AI, using it to greatest benefit, or are chemists on the verge of being replaced, too?
Laboratory automation is important in my life. I am not the first generation that employed computers to obtain and analyze data. The venerable DEC MINCs populated the labs at the University of Colorado, Boulder, when I got there. The new computer on the scene, the IBM PC, promised to be the workhorse of the future. There was no Lab View, only early generation IO boards that didn’t have drivers for the compilers of the day. Coding in assembly code, writing drivers for the first generation data acquisition boards was required. I met the woman who would become my wife while teaching her how to write the assembly code drivers she needed for her experiments. As I said, lab automation is important in my life.
I came to Dow and confronted many problems that were solved with automation. I worked on automating water analyses in my first Dow job, replacing hand titrations done by an operator; these were very boring, repetitive hand titrations of waste water. It was not necessarily a job folks clamored for, making it a reasonable job to automate.
In fact, the industry that I entered was already automated. Long gone were the days of the operator controlling a process from a one-legged stool to insure alertness. Control rooms were already computerized. I have observed consolidation as multiple plants are now operated from the same control room. Chemical industry productivity continues to steadily climb as the combined impact of increasing scale and automation lets fewer employees make more products.
There is a stark difference between academic and industrial labs, a difference that was far narrower when I joined industry. Academic labs are still manual operations. Labor is in excess in academic labs and productivity is not a key metric in academic research. As the combinatorial wave began influencing industrial research, I observed strong academic pushback. Edisonian was used as a pejorative, the opposite of rational approaches. I observed first hand Dow’s embracing of high-throughput experimentation, playing a small role in our transformation. The discovery and development of heterogeneous catalysis, the area in which I spent most of my career, embraced faster analytical methods, automation, and even robotics. The number of experiments under the control of an individual chemist increased and the pace of discovery increased as more options could be tested in a given time period. We moved from being severely equipment-limited to having lots of equipment capacity. Gallows humor predicted management’s embrace of automation was a way of reducing employees. However, the resilience of chemistry prevented it.
The fancy equipment hasn’t outrun the creativity of chemists. We weren't idea-limited when we did a reaction a day. We aren't idea- limited when we can run 100 reactions a day. The creativity of chemists keeps us ahead of technology’s march. The fancy tools are just tools, only able to deliver in the hands of a skilled chemist capable of devising the pathways and compositions worth testing.
I attended a Green Chemistry Gordon Conference in 1996 where I first confronted the concept of computer algorithms for design of synthetic pathways. Present a target molecule and the computer would use its database of reactions to provide the best route. Routes could be optimized for the parameter, or set of parameters, of choice. Cost or fewest steps were the obvious choices, but the algorithm could also be taught the Twelve Principles of Green Chemistry. Once taught, it could optimize on Green Chemistry.
A talk by James Hendrickson remains a memory after many years. The memorable part of the talk came in response to a question about successful application of the code for route selection. The answer was sublime. The algorithm was being used, but Dr. Hendrickson reasoned, no one was admitting it. He reasoned that chemists were recognized for their creativity in finding better routes to known molecules or new routes to new molecules. Ceding this important role would devalue the chemist using the code and, ultimately all chemists. No one was willing to discuss successes meaning that there had to be successes.
Over 20 years has elapsed and I have not observed algorithms taking over. Chemists still find employment. Great synthetic chemists are still in demand, as are formulation chemists with great hands. It could be that IBM’s Watson just hasn't gotten around to chemistry yet and that the moment when machines replace chemists will soon be upon us. I think not. One of my main reasons for optimism is the Tricorder. Even in the future when the miraculous Tricorder is standard equipment, there will still be a Science Officer. Machines can't do it all. In my vision of the future, chemists will live long and prosper.
Thompson, Clive; "Rage Against the Machines", Smithsonian, January/February 2017, page 21-25. [www.smithsonianmag.com/innovation/when-robots-take-jobs-remember-luddites-180961 423/#TY8VWDSdgsLcJhKy.99]
Frey, C.B. and Osborne, M.A., 2017. The future of employment: how susceptible are jobs to computerization? Technological Forecasting and Social Change, 114, pp.254-280.
National Intelligence Council; "Global Trends: Paradox of Progress", January 2017.
Chabris, Christopher; "Computers Get Closer to Mastering Poker", The Wall Street Journal, 4 March 2017.
Mark Jones is Executive External Strategy and Communications Fellow at Dow Chemical since September 2011. He spent most of his career developing catalytic processes after joining Dow in 1990. He received his Ph.D. in Physical Chemistry at the University of Colorado-Boulder doing research unlikely to lead to an industrial career and totally unrelated to his current responsibilities.