The Ethics of Analytics Systems

  1. Introduction
  2. Topics
    1. ¶Indentity
    2. ¶Privacy
    3. ¶Ownership
    4. ¶Reputation
  3. We Need A North Star
    1. ¶Company Values & Policies
    2. ¶What About “Corporate Social Responsibility”

Introduction

The “Illuminati” is an interesting metaphor to use when comparing how people feel about organizations that might know a little too much 4. There mere possibilities of what the data might be used for tends to drive some people mad. Whether you are a conversationalist on the Illuminati or not - the thought is still very real that most people get a little timid about the thought of what an “all-seeing” organization COULD do.

As recent years have rolled out, a ton of convenience has been offered in the way of websites, mobile phone apps, and other incredible innovations. What the public fundamentally does not understand - is that your phone or your browser is a whispering traitor to you every day. Your browser leaks information on you every day to various websites. Sometimes it just t to seemingly benign facebook. The goal here is not to scare people into a fear-mongered knee jerk, but to shine a light on exactly how easy it is to become a big brother company.

Companies operating “big data systems” tend to have an incredible potential to empower mankind, but left unchecked the responsibility accompanying those systems is often ignored. In these large companies, we need to have more of a culture of asking “even though we can - but should we”

Employees are increasingly caught in the cross hairs of an “interesting idea”. Without a plan in place or at least company guidelines, the opportunity will speak loudest in the language of glory or greed. Knowing your boundaries is incredibly important as a data collecting company. It helps employees weigh out the creepy factor with the best-case benefits in real-time. “Best-Case benefits” meaning we don’t often give equal credence to each feature we think of the coolest one and that mobilizes the rest of our talking points. With the lack of prior planning, or guidelines each employee reverts to her own moral code - where things typically end in agreeing to disagree since the distribution of people getting creeped out over technology has an incredible wingspan. Worse yet, for many opportunities, there is not an obvious creepy factor risk, and on these projects, we often fall into the pitfall of unintended consequences.

It is useful to call out that “big data systems” are ethically neutral. Only until an agent imposes an agenda does the system become mobilized for a cause. Furthermore, the discussion of these causes comes with funny and polarizing opinions. There are “good causes” and “evil causes” Needless to say these causes get compared against company values quickly. And while many companies have some platitudes with rowing teams hung on the wall. Many frontline workers understand the implied values. Most company values are implicit in the business: happy customers, fair working environment, etc.

Consider an example of this where a company like “BlockFlustered Videos” kept records of customer videos rental transactions. That customer log is JUST DATA until someone notices your data and calls into question your character as result of your rental history. True Story. See the case of Robert Bork who was nominated to the US Supreme Court in 1987. He also was taken to task during a Senate confirmation hearing about his “questionable” rental log. 5 Within the next year the Senate, in an effort to protect their own dirty movie rental history declared an individual video rental history a private matter see the Video Privacy Protection Act

This example depicts the fact we like to think of our politicians in a positive until the data allows us to do so - it also shows the power of data to turn our opinions on a dime. Consider if Rosa Parks was on Twitter, JFK had a blog, or Churchill has posted selfies to Instagram, would we still think so highly of these public figures. At best, life gets a lot more complicated where these figures who often enjoy these high ivory pedestals that people place them on - become mired in everyday-ness. What will our grandchildren know about us that we never knew about our parents or grandparents?? Perhaps the better question is - Is this a good thing?

Thus Employees have the hard job of navigating the fog of fame and glory within the opportunities for innovation, deeper insights, broader outlooks on customers, and better customer engagement. We try to balance this against the risks: security, privacy, legal compliance, customer Engagement

In terms of the categorical types of data, we tend to collect there are 4 main issues with correlating disparate data sets.

**Four Key Topics: **

  • Identity
  • Privacy
  • Ownership
  • Reputation

Topics

Indentity

All we can definitely say is that our identity is a moving target that moves with how we chose to define ourselves. This changes over time. Your identity used to be closely tied to the type of credit card you had, recently it has moved to cell phones, and then beyond the hardware and into profiles that we would create about ourselves online. These profiles differ based on the ecosystem - and the activity that we generate changes over time as well. So what we can say the semblance that we like to put forward is complicated.

Consider the fact that Zuckerberg has declared maintaining multiple identity variants is a lack of integrity 11; whereas Chris Poole (founder of 4chan) argues identity is prismatic. 7 and yet Fred Wilson 8, a very successful VC is famed for saying that we need some lightweight notion of identity. Effectively meaning some rough notion of who you are, done in a way where the person projects a curated view of themselves into public cyberspace).

What types of identity are we collecting? Cookie Id, Device Id, perhaps the user gives their email address, or Lat/Lon to check in at a physical location. The question is do any of these identity types stand as a risk to the company holding that information - clearly, there is a value to collecting and analyzing the data - but has anyone ever couched the discussion around calculating risks too?

Privacy

Consider the bone-chilling fact that 87% of all Americans can be identified by their Gender, DOB, & Zip. So consider that the next time you have to put that into your favorite “Crazy Coupon App”. There is a wonderful axiom by a Law Professor at Colorado - Paul Ohm.

Data can either be useful or perfectly anonymous but never both - Paul Ohm

With non-anonymized data, again the question is what we intend to do with on behalf of our users.

  • Is it clear that we would do this?
  • Is it clear that we should do this?
  • Is this feature within our charter as what our customers expect us to do?

Ownership

  • What does it mean to own data about ourselves
    • In an offline world do I own the fact that I am 6: ft tall? Probably not!
    • In an online world do I own that? It depends on where you said it and using what platform.

Reputation

  • Your reputation used to be determined by people who knew you + maybe their circle
  • Nowadays someone on the other side of the world can determine I am reputable or otherwise solely based on an identity that I may or may not have been involved in.

We Need A North Star

One of the key questions to grapple with is - not only “Would my users want me to do this for them?” - but acknowledging that you have a range or a distribution of users. And not just considering your convenience driven users but giving thoughtful consideration to your privacy oriented users as well.

Company Values & Policies

Are ultimate what shape how employees think about this.

What About “Corporate Social Responsibility”

CSR was a buzz word a while back. Where did it go?
We need to evolve our user personas to include things about how our users feel when they see things online that impart a platform has bought or sold her data recently.

Make your data handling actions not something needing to be hidden or even obscured. But something you could put on a video and be proud of.

References:

essay
people process data ethics

Comments: