Would you trust a physician who has never taken the Hippocratic Oath? What about an attorney not recognized by their local Bar Association? Probably not. We like to think that our medical and legal professionals have our best interests at heart, and that there are traditions and institutions that help to ensure ethical responsibility.
The same should be true of the professional cadre who operates and defends analytics products. CAP accreditation represents the first serious foray into the world of ethical standards for analytics professionals. While INFORMS makes the code available as a guide of conduct to the general membership, CAPs are required to “take the oath,” so to speak, to obtain their certification.
Although the ethics code for CAPs does not yet evoke the kind gravitas associated with other 1,000-year-old professions, its emerging importance is a product of the digital age. As technology develops, the tentacles of analytics applications will extend to many facets of our daily lives. For the present, however, most of us will experience the implications of ethical conduct as a sense of trust with two interrelated facets: trust in the results of our analytic products and trust that these products have been ethically developed.
Examples of how trust in results matters crop up in new places all the time. Most recently, the popular press has been fascinated with the journey of IBM’s Watson into the doctor’s office. Watson marketing efforts and even research projects describe Watson’s ability to make targeted recommendations for cancer treatment. An interesting question arises: should Watson take the Hippocratic Oath? Presumably the doctor remains accountable for ethical conduct in the medical, but this example highlights the importance of being able to trust the analyst who created your product. Far from being a back-room function, in our digital age, our trust in analytics progress will either be an impediment or a strategic asset. So far, the results are not encouraging. Industry standards and credentials such as the CAP can be an important signal in trustworthiness of outputs, and businesses and public organizations should be savvy enough to insist on them in the future.
Trust that analytics have been ethically developed has broader implications for society than a single product, though businesses that fail to take heed can suffer the consequences of a bruised public image. This facet of trust is more about whether our algorithms promote harmful habits, propagate undesirable behaviors (and potentially latent values), or otherwise harm their intended users. Facebook’s recent journey into the 2016 election touches on all three of these. Again, rather than being a back-room function far from the attention of the public, analytics were thrust into the forefront of the public eye as we all questioned what our feeds were really telling us and why. Society learned that analytics can create an echo chamber for hate speech, propagate foreign influence, and be manipulated by businesses.
Both the IBM Watson and Facebook examples highlight how trust in analytics will only become more of a vital interest in the future. Imagine for a moment, Mark Zuckerberg in his congressional testimony telling the team that his analytics team holds themselves to the highest standards of ethical conduct and followed the guidelines set out in a code of conduct. Now imagine that those same analysts took this code seriously and were held professionally accountable to it by their peers. I am sure that the fallout would not have been lessened, but I might sleep a little better knowing that there is a standard being followed as part of a broader community. It may not be a Hippocratic Oath yet, but it is definitely a start.