Seek and You Shall Be Found

by Ethics Newsline editor Carl Hausman

Here’s a scenario: Candidate A, a Republican, is under attack by Candidate B, a Democrat. Candidate B is claiming that Candidate A will cut funding for mammograms. The attack is aimed at diverting support from Republican women, viewed as a vulnerable segment of Candidate A’s likely voter pool. Candidate A counterattacks by distributing a video ad that describes his mother’s ordeal with breast cancer and accuses his opponent of using scare tactics.

The wrinkle: It’s an internet pop-up video ad, and it appears only on the screens of women identified as Republicans who have searched for information about breast cancer.

If this scenario makes you uneasy about the mixture of politics and internet data collection as well as concerned about what might happen when such a tactic becomes possible, I must note that your concerns are so three years ago. What I described is an actual ad used by then-candidate Chris Christie, a Republican, to rebut campaign claims by then-incumbent New Jersey Gov. Jon Corzine, a Democrat.

As reporter Tanzina Vega recounted in an excellent article in the February 20 New York Times, “microtargeted” ads are part of a revolution that sees digital advertising currently accounting for about 15 percent of total campaign spending. The technique is being utilized by the current crop of GOP presidential candidates, including Mitt Romney, who produced separate ads for those recipients who statistical analysts determined are not aligned yet with a political candidate (for them, an image ad, designed to stress Romney’s likeability and family orientation) and another aimed at likely supporters (in that case crafted with a much more strident message and urging them to be certain to go to the polls).

There are two major technologies behind microtargeting, both highly developed but generally not blinking too brightly on the public’s radar.

First is search and keyword analytics, the process of tracking a surfer’s searches and vocabulary and serving up ads based on what the search engine’s algorithm determines are the user’s exploitable interests.

Try it yourself: Use Google’s search engine to explore the word “bankruptcy.” You’ll get search results, all right, but also a collection of (clearly identified) paid ads for law firms and other agencies trying to sell you services related to bankruptcy resolution.

You’ll notice the same phenomenon in the ads that appear next to your email in many web-based email programs. Key words are identified and likely ads summoned based on an analysis of your messages’ content. (The technology isn’t perfect: When I was a member of a local municipal board and made an unpopular decision, I received a heated email from a constituent who accused me of being a “wolf in sheep’s clothing,” a message charmingly accompanied by ads for wool-washing detergents.)

Second is the science of predictive analytics. Number-crunchers are becoming expert at prognosticating future habits from past actions and can uncannily utilize aggregated data, such as purchase habits, to assess your buying predilections. Do purchases on your store card indicate that you’ve bought a pregnancy test and vitamin supplements? Be prepared for on onslaught of ads for diapers and baby food, because you are an ideal customer — reachable and persuadable before you actually start shopping for the product.

Discussion of the ethical implications of data collection is nothing new; Newsline generally features a story each week on online privacy. But the emerging trend of microtargeted political advertising raises three unique ethical questions:

1. Is it right for a candidate to fragment a political message into so many pieces that it essentially can tell anyone what they want to see and hear (at least according to the algorithms)? We’re all aware of the scenarios popular in political comedies in which the candidate promises one thing at the first whistle-stop but makes precisely the opposite promise 50 miles down the tracks. On occasion, an enterprising reporter along for the whole ride would catch the double-talker in the act, but it’s unclear if the ephemeral nature of pop-up ads lends them to reliable scrutiny for consistency. Politics fashioned in an echo chamber is assuredly not a promising addition to enlightened political discourse.
2. Is there moral justification for the intrusive mining of data and coupling the process to persuading voters? It’s one thing to hawk small-breed dog food to someone who buys a Pekingese at the pet store and discloses the information via a store “rewards” card. It’s perhaps quite another to use voter registration information coupled with profile data based on web pages viewed, publications read, and charitable contributions made. One can make a reasoned argument that public policy decisions have an eminently more profound implication than consumer purchases and ought not to be treated as purely commercial commodities.
3. Does targeted identification of valuable or malleable voters exclude some from the process? Vega’s piece points out that some critics contend that microtargeted advertising based on composite data could exclude those deemed to be members of a less desirable or less persuadable demographic, effectively “redlining” certain groups out of a portion of the political process.

As is often the case with technological innovation, the technology gallops ahead while the law lags behind, and even if the law seemingly catches up, the technological gazelle usually has darted in an entirely different direction. Ethics, responsible judgment, and informed scrutiny, therefore, are probably the only viable tools for heading off what could be a cyber-unbalancing of an electoral process that already is badly out of kilter.

©2012 Institute for Global Ethics