Technology is Too Complicated

Last May, a Tesla Motors self-driving vehicle’s sensor malfunctioned when it failed to differentiate between a white tractor-trailer and the bright sky, causing a fatal crash. Two months later, a power outage at Delta Airlines' Atlanta hub resulted in the cancellation of thousands of flights, and the loss of more than $100 million. Reports vary on the exact causes for both incidents.  

To abandon testing of automated driving technology or to mothball the airlines' years-old computer systems would be an overreaction. The Delta event is almost certainly an aberration. The autonomous vehicle crash may typify the sobering, sometimes fatal consequences of still-imperfect technology on the cutting edge of human understanding. But these system failures do raise a timely question: Is technology already at the limits of human comprehension? 

Complexity scientist and author Samuel Arbesman thinks so. He has been writing about our “overcomplicated” technology at Aeon, Wired and Nautilus, among others, for years. Mr. Arbesman’s latest book, “Overcomplicated: Technology at the Limits of Comprehension” (Current, 2016), argues that technology is already too complex for the experts. He suggests, however, that by “peeking underneath the hood of technology,” users can inoculate themselves against the unhealthy perspectives bred by human uncertainty about machines. Below is an edited transcript of RealClearBooks’ conversation with Samuel Arbesman. 

Q: What inspired you to write “Overcomplicated”, and what did you learn in the course of researching this book?

I am fascinated by the complexity of systems. Although our technological systems, in general, are becoming more complicated, more interconnected every day, I also found that this phenomenon of increasing complexity is happening to our domestic legal and statutory codes (as lawyer and author Philip K. Howard says, “modern law is too dense to be knowable”), public policy programs, and engineering systems, among many other examples. This question – ‘why do so many human innovations tend to become more interconnected, more overcomplicated’ – was the animating principle of my book. The true extent of the complication was shocking.

Q: You are a “complexity scientist”, which, for the mere fact that such a profession exists, seems to validate your thesis that technology has grown so complicated as to become incomprehensible. In layman’s terms, what is complexity science and why is it important?

Complexity science is a rich and exciting field that touches on a variety of powerful ideas, technical areas of study and mathematical frameworks in a quest to find patterns or meaning in complex systems. To put a definition to it, I say that complexity science is the quantitative study of vast and complicated interconnected systems, ranging from living things (biological organisms) and ecologies to the World Wide Web and even collaborations between film actors (think “Six Degrees of Kevin Bacon,” where you try to connect every actor back to Bacon through film co-appearances). In our era of disaggregated information and rapidly developing technology, complexity scientists are developing innovative methods to address pressing problems in science, government and industry. 

Q: And this notion of increasingly complex, interconnected systems leads you to suggest that our society is caught in the “Entanglement” of over-complication. How did you arrive at this term?

This term denotes a sprawling web of interconnected systems and technologies that are not just complicated, but sometimes inexplicable altogether. The computer scientist Danny Hillis argues, “Our technology has gotten so complex that we no longer can understand it or fully control it. We have entered the Age of Entanglement...Each expert knows a piece of the puzzle, but the big picture is too big to comprehend.” We no longer foresee the problems that will arise when a complex technology interacts with users or with other systems. Compounding this problem is the fact that not even the experts can understand the whole system or its errors.