Although security researchers warn us of the possibility that the internet could go down—a massive DDoS attack, or an attack on the electric grid, could make internet access impossible for many—the prospect of living a normal life without the internet, even for a few days, is increasingly difficult to imagine. Our attention has been firmly grabbed by all the conveniences and possibilities of the internet, and the normative value of that attention is overlooked by developers looking to gain ever more of it. We can make apps more addictive, games more immersive, virtual reality better and more informative than actual reality, and we can outsource more and more of our working memory to Google Maps, Evernote, iCalendar, Gmail, Dropbox, and other internet resources, but is this good? Is a dependence on the internet improving our ability to do work that we value, or is it increasing the flakiness of our attention through multitasking?
Byron Reeves doesn’t evaluate whether students who switch screens every minute (or every twelve seconds) while writing an essay are writing better or worse essays than students who stay on one screen until the composition is complete. Other researchers have determined that task-switching and multitasking make us worse at certain forms of mental activity (driving a car, for example, or remembering and adding numbers, or carrying a conversation). Some of the most interesting conclusions from Professor Reeves’ research are how we actually arrive at a conclusion: a finished essay, a completed online purchase, an email sent. The amount of work—and distraction—necessary to get through our daily lives is astonishing, and would be of great interest to retailers, entertainment providers, and teachers alike.
Tim Wu has taken the opposite position, and assigns strong normative values to the actions that “attention merchants” take to grab our attention. The advertising model, to Professor Wu, is a lose-lose situation for publishers, consumers, and even advertisers themselves, because the ad model rewards click-bait and makes independent publishing difficult. I can't get too worked up over internet advertising being evil in and of itself (even though I use three adblockers simultaneously, which sometimes crashes websites in very weird ways) because I would much rather that everyone be able to access the New York Times or the New Yorker rather than a paywall system that assumes high-quality journalism is only for those who can pay for it. I agree with Professor Wu that tracking and surveillance curbs certain search behavior, and provides most of the annoyance around ads (e.g., you search for airfare once and see travel-related ads for weeks). But I don't think the answer is more of a privacy shield against ads knowing anything about us—what Professor Wu describes as the "deal", the attitude that ”I’ll eat my advertising vegetables if I can watch my sports cake because the sports cake is that good”. Does one participant in this bargain have to lose? If there are products that we actually would want to know about (e.g., a new book by your favorite author, a new item of clothing with a cute fox print if the ad server knows you like cute animal prints) and that would be enjoyable to know about, then the problem is that advertising still uses a very elementary level of inference. This ad model says "she searched for vacuum cleaners last week, let me show her more vacuum cleaners" and has no way of knowing whether she already bought the vacuum cleaner, or was searching for information about the vacuum cleaner she already owns, or that she is extremely loyal to one brand of vacuum and would never switch. Perhaps we need a better AI coordinating advertising to be able to serve ads with that level of specificity, and we're still on our way there.
The media and entertainment market is deeply entwined with the advertising business; have they figured out a better way? Find out next week.
Listen on iTunes, Soundcloud, or at http://worldview.stanford.edu/raw-data