Some professions are in positions of public trust, even though many of their practitioners are in the private sector and may have no compunction about amassing vast wealth from their efforts. For any given society, you can identify these professions in any of several ways: does the government require rigorous education, testing, certification, and licensing before people are allowed to practice the profession? Have the practitioners established associations that codify, disseminate, and enforce professional codes of conduct? Do practitioners put their careers in jeopardy if they fail to conform to these codes of conduct? And so forth.
Doctor, lawyer, and certified public accountant are examples of trust-based professions falling within these core criteria. Recently, there has been some debate in the industry whether data science should also be considered a profession in position of public trust. Considering that the general public is starting to invest considerable trust – and in some cases, distrust – in what data scientists do for a living, this is an important issue has been raised: should there be a professional code of conduct for data scientists?
One way to back into an answer for this is to ask whether there should be a code of conduct for scientists, in general. After all, scientists are respected, authoritative, trusted practitioners of the art of acquiring knowledge from empirical investigations. Or, if a code of conduct that spans all scientific disciplines seems too broad to be useful, maybe each branch of science should have its own code of conduct. For example, social and biological scientists must respect their subject matter (i.e., their fellow human beings, other sentient life forms) in a very different way from the physical sciences (e.g., a lump of plutonium feels no pain).
If we limit the discussion to data scientists whose work has direct or indirect impacts on individual people or the society at large, there’s a credible case to be made for a code of conduct. Just as society has every interest in keeping quack physicians from practicing their bogus trade, we should be vigilant against potentially unscrupulous data scientists who deceive their employers and customers, invade people’s privacy with wanton glee, and so forth.
Even if it has no legal force, a code of conduct that has the clout to keep malfeasant data scientists out of this line of work is a great idea. To that end, I recently read an excellent article that proposes just such a code. Authored by Michael Walker, a Denver-based system integrator, the piece lists the various things that data scientists must intentionally avoid doing in order to earn the public’s continuing trust. Walker provides a comprehensive litany of the frauds, falsehoods, deceptions and manipulations that, if left unchecked and uncorrected, can stain the data scientist profession’s perceived integrity.
If nothing else, every organization that employs data scientists should institute a professional code of conduct of this sort. To give it teeth, the code must be used as a way to ensure that all staff data scientists, contractors, and job seekers know what sorts of conduct might constitute grounds for punishment or dismissal.