Google is facing scrutiny in the wake of the announcement that it has near-omniscience in tracking the purchases of Americans. The search — I mean, A.I. — giant made the revelation on its blog Tuesday.
Google claimed that between new partnerships and a “revolutionary” MIT-inspired approach to data modeling, it can scrape and analyze up to 70 percent of all credit and debit card transactions across America. Since 2014, that’s over 5 billion transactions.
This is an awkward moment for Google. On the one hand, its whole business is selling advertising space and, sensibly, it wants the people buying that space to know its true value, so they’ll pay the most for it. It turns out that Google has spent years building the ultimate system for tracking people’s purchases virtually wherever they go, allowing ads to be perfectly targeted to each person’s individual habits — but what good does that do for Google, if it can’t mention that fact to the people buying the ads, aka, the public?
That’s the conundrum Google has faced with its purchase tracking software: how does it get the word out about what it can do, without getting the word out about what it can do?
“What’s really fascinating to me is that as the companies become increasingly intrusive in terms of their data collection, they also become more secretive,” Marc Rotenberg, executive director of the Electronic Privacy Information Center, wondered to the Post.
The main concern, obviously, was privacy: how is this data being collected, by whom, in what ways?
The answers are still unknown, as the company has delivered firm no-comments across the board. It has said that it only purchases data that the seller has the “right” to collect and sell, but it did not confirm that this mean that the data had been collected with consent. Most customer loyalty programs have a tracking provision that members opt into, but such programs don’t account for 70% of all card-based purchases.
Google is Keeping Mum
Google is also keeping quiet about the details of just how it crunches this data once it has it, but it did say that to protect privacy it has developed a “double blind” approach to modelling that’s based off of this 2011 paper commissioned from MIT researchers by Google and Citigroup. It allows computation to be done on data that’s still encrypted, meaning that it can correlate one person’s encrypted personal information in one database with that same person’s also-encrypted purchase data in another database, without ever decrypting either one.
In principle, Google never really “gets” the data, just gets access to it long enough to do this double-blind check.
Everything Remains Anonymous
The idea is that this ensures that everything remains anonymous — which makes Google’s reticence a bit confusing. The company has a long history of being vilified in the press for wanting to study people’s behavior, but its tight-lipped attitude is not helping. If it wants to avoid seriously souring the media narrative, Google needs to center its consumer tracking business around the privacy of customers, and not its proprietary technologies.