Inside Story

Ethics, O.R. & analytics

Peter Horner, editor
peter.horner@mail.informs.org

Anyone with a conscience believes in ethics – in their personal life, in their business life, in their any other life. The definition of ethics varies depending on the dictionary, but basically it boils down to doing the right thing, even if that “thing” might cost you your job, money, prestige, friendships, whatever you may hold dear.

When its Certified Analytics Professional (CAP®) program was launched several years ago, INFORMS included an ethics component that outlined norms of behavior that CAP recipients were expected to uphold. In 2016, INFORMS issued a set of ethics guidelines for those working in the O.R. and analytics field to “aspire to.” INFORMS members David Hunt and Scott Nestler spearheaded the creation of an INFORMS Ethics in O.R. & Analytics Group to promote the guidelines.

“Work done by members of INFORMS increasingly impacts peoples’ lives as the use of our algorithms and models spreads and as data collection becomes more detailed and more personal,” Hunt and Nestler wrote in an article for OR/MS Today. “Yet too often we become absorbed by the elegance of our models and research, failing to consider any potential ethical implications. Or, we find ourselves faced with a situation that seems unethical, and are forced to determine the best course of action without previously having established exactly what we consider to be ethical.”

In this month’s lead feature (“When good A.I. goes bad,” page 20), contributing author Joseph Byrum brings up ethics as it relates to artificial intelligence and machine learning, two of the hottest topics in the O.R. and analytics orbit right now. Writes Byrum: “The lesson as we develop increasingly powerful A.I. solutions is that the A.I. can never substitute for due diligence. The more we turn over decisions to A.I., the more we must ensure that procedures are in place to catch the mistakes made not by the A.I., but the humans who set up and operate the systems.” In other words, Byrum says, validation and verification remain crucial.

The feature lineup is backed by our biennial survey of vehicle routing software (page 42). Given the widespread media coverage of driverless vehicles and delivery drones, I thought it would be insightful to ask survey respondents what impact, if any, drones and driverless vehicles would have on VR software in the future. Here are a few of their responses:

  • “Planning will probably not change, but more intelligent systems will need to be created to interact with them for execution.”
    - Christian Lafrance, Clear Destination
  • “If drones and driverless vehicles deliveries do become mass market, the routing problem is essentially the same but will likely require new constraints and customizations.”
    - Phil Welch, Open Door Logistics
  • “Drones and driverless vehicles will add new constraints to existing systems, but their impact will be minimal. Other future changes will have much bigger impact.”
    - Felipe Carvalho, Widescope
  • “Driverless technology will be subject to the same constraints in the transport network as a human driver. Good savings opportunities come with shifting solutions in the time domain. The main advantage is that drive-time constraints and other human constraint factors can be improved. For driverless vehicles, existing algorithms would be able to find new optimizations that may be infeasible at the moment.”
    - Ruben Filter, Intelligent Routin