Data Ethics in Corporate Cultures: How do we encourage ethical behavior?

Most corporate ethics policies fail, because they require individuals to be brave. They ask people to jump off a cliff and so rarely provide a parachute.

Data Ethics in Corporate Cultures: How do we encourage ethical behavior?

As data ethics enters the mainstream, more people are looking for ways to incorporate them into their culture. Some look at developing policy or even, dare I say it, their own code of ethics?

Before we run too far with this, let’s review what we already know:

  1. Ethics sit in a fuzzy space between morals and laws. We rely on them when no laws exist to guide us and the future is uncertain.
  2. Ethical decision-making works best with an established framework.
  3. Ethics are a hallmark of a practice profession where knowledge workers often have more information about ramifications than decision-makers.
  4. Specifically, data ethics encompass far more than privacy and accuracy, and often make us uncomfortable.
  5. Data ethics also requires us to recognize “raw data” is a misnomer and that we are actively involved in the decision-making process from start to finish.
  6. Ethics require us analysts to look at our work differently (such as when we visualize sensitive topics such as suicide (content warning) as we can do harm.

Now, corporate friends, let’s take this on!

Data Ethics and Culture

One of the scariest things about ethics is that they often rely on bravery. This seems like a small problem compared to what ethics often ferret out, but society rarely rewards the brave in that moment. Standing up for something often comes at a cost, hence why so few do it. Only in retrospect do we often see and honor courage. Most times, it’s easier to go with the current, rather than against it. Those that call out ethical dilemmas or alternate approaches face a struggle – or worse, ramifications – for daring to call attention to a problem.

Your data lorax, if you will, speaks for the data, for the data has no tongues. Unbeknownst to so many, your data is often people, hidden as rows of numbers or as artifacts of behavior. In business, we use this, make decisions, and affect certain worlds. Historical ethical thought has long contemplated this dynamic. Nicomachean Ethics, by Aristotle, still drives many ethical discussions today, which distills frequently to one main question: what rights do we/I/they have to do X? This question also delves heavily into another, what makes a good life? Some circles call this eudaimonia.

To incorporate data ethics in our culture means recognizing the following uncomfortable truths, or sticky wickets:

  • Data shapes much of the world today. The analyses and conclusions we make have a direct effect on someone else. Sticky wicket 1: Our work does not exist in a vacuum.
  • Data loraxes are often created by adversity. They tend to recognize barriers because they have often experienced something systematic in nature blocking them from eudaimonia. Sticky wicket 2: our lorax is likely to be underrepresented in some way (race, gender, religion, illness, socioeconomic background, language…pick several).
  • Ethics are rarely protected in the workforce. Whistleblowers are the extreme end of this. Professional organizations, such as the bar association for law, seek to mitigate some of this and enact some protections. See points earlier about ethics and practice professions. Sticky wicket 3: you can ‘ethics yourself’ out of a job.

Culturally, we understand those who provide view of an ethical dilemma are risking something. What can we do to help?

Codifying Ethics

This is where so many want to start. In companies, we love making signed agreements. I know, I’ve drafted several of them myself (complete with legalese jargon and Latin thrown in for spice). Ergo, we move onward.

When it comes to dealing with data ethics, though, this is not the place to start. For starters, there is no standard unifying body for data workers. (There are definitely starts and potential, but not standard.)

Unifying bodies are awesome when it comes to dealing with ethics. They identify those who qualify under its jurisdiction. They define the skills and education required to truly be functional in the industry, often with fine-lined names. They normalize what practitioners can and can’t do through standards and, best of all for companies, they certify all of the above through a non-tool-specific accreditation process. This is more than proficiency in software or proof of functional skills, but of ethical thought and appropriate decision-making. And yes, with certification comes revocation for problems. For those within the organization, they also help legislate, so bravery becomes less of a requirement.

Most corporate ethics policies fail, because they require individuals to be brave. They ask people to jump off a cliff and so rarely provide a parachute. They mandate behaviors with little contemplation to what causes the concern in the first place, often making loraxes the first, last, and only line of defense against a large societal blunder. Ethics exist because laws fail to keep up with innovation, and helping some populations often risks hurting others without realizing it – the more we keep that in mind, the better we will do when we adopt ethics as part of our ethos.

Embracing Ethics

Our natural tendency is to hate ethics. So often, they tell us we can’t when we really want to do something. Some argue they stifle innovation, ruin fun, and make things too complex. Would you want your doctor saying this about your health, or your lawyer about business protection? Data ethics allow us to create a more just world for most, not just the rare few. Even with the best of ethics, we will still do harm, which is why unity and laws ultimately must come into place. Ethics are not a substitute for long-term needs; they must be codified into law, their protections upheld by the larger society.

When we embrace ethics, we allow a richer dialogue about what we’re doing as a company. It allows us to consider more angles and adjust. It means slowing down just enough to truly see the whole scene, and not just a blur as we run past. Normalize ethical discussions and make them visible. Show their effects.

Name the Values

We fear things the most when they lack a name. Your code of ethics should not be Voldemort, He-Who-Must-Not-Be-Named, or any other variation of Tom Riddle.

Do you want to be fair to your customers? You can safely call this justice. Transparent? We recognize this best in an ethical framework as confidentiality, which is about trust rather than secrecy (root being to be a confidante), and autonomy, which lets people maintain a level of control. We also recognize our ability to do harm and actively work to avoid doing so, which is the real meaning of beneficence. These terms and others (which you can find here) are the basis of codes of ethics across industries.

Write them down. Use these words and others and you explore issues. When you run into a dilemma, it’s this vocabulary that will bring insight and potential solutions. Is this a problem because it intersects with with autonomy, justice, and confidentiality? Naming the values is key.

Providing Parachutes

Do you want someone to report ethical concerns? When its embraced culturally (like innovation or, if you must, “disruption”), it becomes a prized value, one we celebrate. We make it easier to jump out of the plane, because all the backpacks have parachutes. We do even better when everyone has had time to practice, including pulling the cord.

The more we increment ethical conversations, such as encouraging concerns to be voiced (and actually taking the time to explore them) before we start down a path, the better we do as an organization. When we look at our data and recognize the very real human footprint in it, we can see how our company or product might reshape society and who it risks locking out or blocking from a good life. We start to see where we may cause harm, despite doing good for others. Both can be true.

Again, people need to see that the parachutes work. Incorporating ethics verbally, but not showcasing them is the same as showing the backpacks, but never letting people see the parachute deploy. If you’re going to jump out of a plane, you want the parachute to work. Likewise, if you’re going to call attention to a big ethical conundrum, you want to feel safe (see sticky wickets above).

Support Professionalism

You want professional organizations to take some of the management overhead of codifying ethics. Yes, it means you’ll have even more people telling you no. This is actually a good thing. It elevated the medical profession, chemistry, and it will data as well.

In the absence of a unified body, start your own council, but be prepared to shift to following an organized, neutral body at some point. Ensure this council is representative of all stakeholders. If everyone can wear the same clothes or purchase the same haircut, you’re doing it wrong. If the broad majority of this council shares the same pay grade, you’re missing out and doing yourself a disservice. Empower those who are least empowered with a voice – we learned from above, they’re most often a lorax. With that, partner them with someone who has the latitude that can do something about it.

Your council is your guide. They need to be visible, but also approachable. They need to do the hard work of acting as the intermediary between loraxes and those who make decisions. Again, it must work and it needs teeth.

Monitor Yourself

Ethics are practiced. Ideally, they mature and become established. If you’re fighting the same battles or the same types of people are always exiting after a year, it’s time to take a good hard look at your ethical program. Chances are, you’re silencing some of your best loraxes or requiring bravery far too often.

If a core value is always challenged, it’s probably time to do some homework. Find out why. Ethical review requires honesty and the willingness to adjust. You can’t implement ethics appropriately for good marketing. Rather, it will change you ideally, more than you change it.

Make this transparent. Others learn and grow as you model the discussions and results. Track concerns and ensure you’re not solely placating people so you can continue to do whatever you want.