Data is a trending topic right now, and data privacy is one of its trendiest subsets. To wit, Charles Duhigg’s investigative report on Target’s data mining for the New York Times spawned a series of follow-ups, in March, The Atlantic profiled NYU Law professor Helen Nissenbaum and her flow-based privacy framework, and the FTC just published a privacy report endorsing privacy-by-design and the “Do Not Track” button. The demarcation line between what should be public vs private is a dynamic and jagged (some might say gerrymandered) one that depends on a piece of data’s original context vs the contexts in which it is eventually used. It seems perfectly reasonable for Foursquare to publish its users’ locations but less reasonable for a third-party dating application like Girls Around Me to provide these locations, along with Facebook profile photos, to its users. It seems reasonable that an online money management service like Mint serves up ads tailored to users’ credit ratings, but less reasonable that banks determine applicants’ loan rates based on their Facebook friends’ credit ratings.
Because we’re storing and analyzing corporate email, user privacy is something that we have to get right. Of course, an employer’s definition of “right” might be different than the employee’s, so we’ve been trying to figure out a definition that will please both. Companies are legally permitted to access their employees’ email, and usually this manifests in explicit/inappropriate language monitoring. As long as employees are aware of the monitoring, this sort of vocab dinging seems reasonable. But what about sentiment analysis, and the inferred knowledge of employees’ mind states it provides? Invaluable to the company, I think, but potentially detrimental, and sometimes errantly so, to the employee. Does explicit consent justify armchair psychology and any actions that result? Even if employees are fully and duly informed of all monitoring and tracking practices, I’m not sure. Take, for example, Cataphora.
Cataphora is a “behavioral modeling and monitoring” software that analyzes employees’ digital and mobile actions from legal, risk, compliance, HR, and brand management perspectives. The copy on its website doesn’t even try to address employees—there are callouts on its news page to articles with titles like “In Defense of Employer Monitoring,” and “Finding Office Buck-Passers, Heroes, and Shirkers.” If employers are not monitoring employees’ digital activity, Cataphora CEO Elizabeth Charnock argues, they are making themselves vulnerable to leaks, blow-ups, and Youtube frittering-induced productivity slumps. In a blog post entitled “Getting Big Brother Right,” Rick Janowski brought up as a use case an employee on the verge of a breakdown due to non-work-related factors. Cataphora could identify and alert management to the employee’s mental state, allowing them to “provide a safety net for someone who might be prone temporarily to making bad decisions or being less diligent than they normally would be.” Aka remove him from fiscal and legal harm’s way before it’s too late. Ooh, Carnival Cruise is having a flash sale! I hear Alaska’s great this time of year!
You could argue that behavioral mining software is just one of the many new “transparent” office measures, which manifest physically in concepts like open and free range offices (a different desk every day!), and culturally in social enterprise platforms like Yammer, Rypple, and Trello. There’s been a push, lately, to besmirch the traditional office, with its many doors and walls and silos. Which is all very well and fine, but there is a point where public property ends and person begins. Perhaps the central tower is too zoomed in to see it.
Web mining consists of crawling the world wide web, and extracting and processing its structure, usage, or content. This