- What do the Big Data Guidelines Mean for Employers?
- August 15, 2017 | Author: Maciej Lipinski
- Law Firm: Borden Ladner Gervais LLP - Toronto Office
On May 17, 2017, the office of Ontario's Information and Privacy Commissioner released its Big Data Guidelines for organizations that engage in the collection, use and analysis of increasingly large-scale bodies of data.
Questions arising for employers
For employers, the two immediate questions arising from the release of the Guidelines are:
- What is "big data"?
- Why should I care?
The Short Answer
In brief, "big data" refers to the tools for collecting massive amounts of information and then analyzing that information in a manner that enables organizations to make meaningful predictions and decisions. The increased availability of a wide variety of massive data sets and complex analytical software programs to interpret them has served to make “big data” tools into an increasingly important — and perhaps even essential — means for businesses across a wide variety of sectors to remain competitive. At the same time, the wide variety and rapid development of such tools gives rise to concerns over both the quality of their output and the legal boundaries that may be crossed where these tools are applied without constraints.
As for why employers should care: big data tools are valuable for attracting talent, managing productivity, and accurately assessing the effectiveness of one's workforce in increasingly competitive marketplaces. At the same time, employers must take appropriate measures to ensure that the big data tools they invest in are delivering results as promised without crossing legal boundaries that trigger new liabilities. To these ends, the Big Data Guidelines offer numerous valuable points of guidance and direction.
Unpacking the Big Data Guidelines
To employers, big data tools may be viewed as a reliable, standardized and cost effective means for managing the workplace. As the Big Data Guidelines highlight, however, such tools are not always all that they are cracked up to be — and the legal consequences of misusing big data tools may be more than most employers bargained for. In particular, the Big Data Guidelines highlight three key cautionary points for employers seeking to capitalize on the benefits of big data:
- Respect privacy interests;
- Recognize potential discrimination; and
- Remember the limits of prediction.
Respect privacy interests
Employers may be inclined to view information posted to the internet — whether as Tweets, Facebook posts, or blogs — as open, completely public sources of data that can then be collected, used and analyzed for purposes such as determining the suitability of a job applicant or assessing a worker's productivity. As the Big Data Guidelines highlight, however, the view that individuals forfeit all privacy rights to content posted online has become "increasingly problematic". This applies particularly in view of big data tools' capacity for synthesizing individual pieces of such online information to uncover patterns that an individual may never have intended to share publicly. Accordingly, in light of the capacities of big data, a privacy interest applies to information based on both what the information means, and what it could mean.
While Ontario does not yet have legislation that generally protects individuals' rights to privacy, its courts have nevertheless recognized intrusion upon seclusion as a compensable tort arising from violations of individuals' privacy rights. In Jones v Tsige, the Ontario Court of Appeal determined that such invasions give rise to liability where they are intentional or simply reckless to the privacy interests of individuals concerned. As increasingly complex big data tools allow users to delve further into individuals' private lives through publicly disclosed information, employers who make use of such tools must be mindful to ensure that these uses do not cross the line between productive application of such technologies and reckless invasion of individuals' privacy.
Recognize potential discrimination
Employers are also likely to take an interest in the ability of big data tools to track patterns over time. This may be seen as particularly useful for employers interested in tracking employees' productivity over time — or even tying such metrics to decisions on remuneration and bonus calculations. While recognizing such valuable applications of big data tools, the Big Data Guidelines also illustrate potential risks that arise from "discriminatory proxies" — a condition that results when tracked variables appear to be neutral but nevertheless closely correlate with a protected ground of discrimination , such as sex, age, gender, or ethnic background. For instance, where the composition of an employer's different departments or locations significantly and systematically varies based on such grounds, tracking productivity differences between locations or between departments and awarding compensation accordingly may amount to discrimination on those grounds.
In Ontario, as in other Canadian provinces, human rights legislation extends protection from discrimination to both current and prospective employees with respect to various protected grounds. Accordingly, employers who use big data tools to make decisions with respect to their workforce — including decisions regarding hiring, promotion, compensation, termination and layoff — must ensure that the data points used to distinguish employees do not operate as "proxies" for one or more protected grounds. At a minimum, satisfying this standard requires employers to maintain familiarity with the statistics and operations used by their big data tools before deploying these tools to make substantive, decision-driving distinctions.
Remember the limits of prediction
The feature of big data tools that is arguably most appealing to users, including employers, is the purported ability of these tools to make predictions based on patterns in the data they collect. For example, an employee's likelihood of succeeding in a position of greater responsibility may be determined by a big data tool that compares certain characteristics of that employee to the characteristics of other employees who have been identified as top performers in similar roles. In view of the appeal of such features, the Big Data Guidelines address the potential for missteps in “profiling” individuals in this fashion, highlighting the potential for false predictions where there is a lack of transparency in the assumptions that go into making predictions in the first place. For employers, this means that one of the main purported advantages of big data tools has potential to become one of their greatest weaknesses by tying the workplace and the hands of managers to software that fails to perform as promised. This potential downside is worsened by the fact that a tool's inaccurate predictions may be difficult to detect for a long period where a user is not familiar with the assumptions underlying those predictions.
Employers are therefore well-advised to be sceptical shoppers when it comes to the market for big data tools and should engage suppliers who are (i) willing to disclose the central assumptions underlying the efficacy of predictions made by their tools; and (ii) able to provide means for this efficacy to be tested when it comes to the employer's actual use of the product.
At the same time, employers are also well-advised to recognize the inherent limits of big data. These tools are usually designed to predict what is true on average — which makes sense when talking about groups of people, but may have limited application to any particular individual. In other words, for the foreseeable future, assessments and decisions concerning individual employees will likely be best left to the discretion of fellow employees, and big data tools that purport to supplant this function should be carefully scrutinized before an investment is made to acquire them.
ConclusionThe Big Data Guidelines indicate that further legal developments governing the use of big data tools in relation to the protection of privacy may be on the horizon. Jurisdictions outside of Ontario, such as British Columbia, have already established laws concerning the protection of privacy that are applicable between employers and employees, and provinces lacking similar legislation may soon contemplate putting such measures into place. Accordingly, the cautionary advice and recommendations contained in the Big Data Guidelines provide employers with a valuable opportunity to ensure that their uses of big data tools fit within the four corners of the law as it currently stands, and that employers are ready to comply with future developments in privacy law and regulation of big data technologies.