December 11, 2013
We like to believe that every visit to Google is a search for knowledge, or, at least, useful information. Sure, but it’s also an act of narcissism.
Each time we retrieve search results, we pull out a virtual mirror that reflects who we are in the Web world. It’s what Eli Pariser aptly described as the “filter bubble” in his 2011 book, The Filter Bubble: What the Internet Is Hiding From You.
Pariser laid out the thinking behind algorithmic personalization. By meticulously tracking our every click, Google–and now Facebook and more and more other websites–can, based on past behavior, make pretty good guesses about what we want to know. This means that two people doing exactly the same search can end up with very different results.
We’re fed what we seem to want, and since we’re more likely to click on stuff within our comfort zone–including ads–Google, and others, are motivated to keep sharpening their targeting. As a result, the bubbles we live in are shrinking.
There’s a price for all this precision, as Pariser pointed out in an interview with Brain Pickings’ Maria Popova:
“Personalization is sort of privacy turned inside out: it’s not the problem of controlling what the world knows about you, it’s the problem of what you get to see of the world.”
The bigger picture
So we’re trapped in a maze of our own making, right?
Not necessarily, thanks to a team of scientists who say they may have come up with a way to escape the constraints of algorithms. As the MIT Technology Review reported recently, Eduardo Graells-Garrido at the Universitat Pompeu Fabra in Barcelona and Mounia Lalmas and Daniel Quercia at Yahoo Labs have developed what they call a “recommendation engine,” designed to expose people to opposing views.
One key, say the researchers, is that those views come from people with whom we share other interests. That seems to make us more receptive to opinions we’d otherwise likely dismiss as folly. The other is to present opposing views in a visual way that makes them feel less foreign.
To that end, the scientists used the model of a word cloud, which allowed study participants both to see what subjects they tended to tweet about most often, and also to have access to–in a visually engaging way–content from others whose own word clouds mentioned many of the same topics.
But what if some of that content reflected a very different political view? Would people instinctively reject it?
To put their theory to a proper test, the researchers connected people on opposite sides of an issue that evokes deeply personal feelings–abortion. They focused on thousands of active Twitter users in Chile who had included hashtags such as #prolife and #prochoice in their tweets, creating word clouds for them based on terms they used most frequently.
Then, they provided study participants with tweets from people who had many of the same terms in their word clouds, but who also held the opposite view on abortion. The researchers found that because people seemed to feel a connection to those who had similar word clouds, they were more interested in their comments. And that tended to expose them to a much wider range of opinions and ideas than they would have otherwise experienced.
In short, the researchers used what people had in common to make them more open to discussing ways in which they differed. They had, their paper concluded, found “an indirect way to connect dissimilar people.”
So, there’s hope yet.
Madness to the method
Here are other recent developments in the sometimes bizarre world of algorithms.
- Nothing like automated “Warm personal regards”: This was probably inevitable. Google has just received a patent for software that would keep such close track of your social media behavior that it will be able to provide you with a choice of possible reactions to whatever comments or queries come your way on Facebook or Twitter. If, for instance, a friend gets a new job, the software would suggest a response, presumably something such as “Congratulations.” That’s right, you wouldn’t have to waste any of your brain power. The algorithm will do it for you.
- Phone it in: Researchers at the University of Helsinki have developed algorithms for determining how people get around--walking, driving or taking the bus or subway–by tracking the accelerometer signals of their cell phones. That allows them to analyze the frequency of their stops and starts. The researchers say it could be a powerful tool in helping planners understand how people move around in their cities.
- All the news that fits: Facebook has tweaked its “news feed” algorithms so that more actual news will start showing up there. The idea is to give greater exposure to links to articles from news organizations on Facebook feeds–which will help make the social media giant more relevant to what’s going on in the world besides friends’ birthdays. The speculation is that this is an effort by Facebook to challenge Twitter’s dominance in generating buzz around current events.
- What does she have to say about the Chicago Cubs?: An Israeli computer scientist has created an algorithm that can analyze huge volumes of electronic data about past events from sources as diverse as the New York Times’ archive to Twitter feeds and predict what might happen in the future. Most notably, the scientist, named Kira Radinsky, has used her system to predict the first cholera epidemic in Cuba in many decades and the protests leading up to the Arab Spring.
Video bonus: Here’s the TED talk that made Eli Pariser and his concept of the filter bubble famous.
Video bonus bonus: There are algorithms for everything these days and, to believe Sheldon, of “The “Big Bang Theory,” that includes making friends.
More from Smithsonian.com
July 11, 2013
“There is no doubt that over time, people are going to rely less and less on passwords. People use the same password on different systems, they write them down and they really don’t meet the challenge for anything you want to secure.”
None other than Bill Gates said this. Back in 2004.
People in the business of keeping data secure will tell you that passwords should have gone the way of dial-up Internet by now. Sure, back in the day, when we only needed them for two or three websites and hackers weren’t nearly so diabolical, we could get away with using the same “123456″ password for everything, without worrying that someone on the other side of the world was a click away from emptying our bank accounts.
Ah, sweet innocence. Now, we have an average of 24 different online accounts, for which we use at least six different passwords. And we need them for tablets and smartphones, too. If we’ve heeded the security gods—although most of us haven’t—we’ve abandoned the memorably quaint for strange, long combos of numbers, letters—capital and lower case—and symbols that dare to be remembered. (Then again, most of us don’t seem to have a knack for this passwords thing, considering that year after year, the world’s most popular password is still the word “password.”)
Not that conjuring up the perfect password guarantees immunity from code crackers. Just last week the giant game company Ubisoft admitted that its database had been breached and advised those with Ubisoft accounts to change their passwords immediately. Last summer’s big cybersecurity caper was a hack of LinkedIn, in which more than 6 million encrypted passwords were exposed.
It’s time, it would seem, for a better idea.
So, who figures to make the first big splash in the post-password world? Right now, a lot of the betting is on Apple, with speculation that the killer feature of the iPhone 5S coming out later this year will be a fingerprint scanner, perhaps embedded under the home button. Some Apple watchers think the iWatch, also expected on the market by the end of 2013, will likewise come with scanner capabilities that would allow the device to verify the user’s identity. Apple tipped its hand last year when it paid $356 million for AuthenTec, a company that develops fingerprint scanners.
Other big names pushing for the password’s demise are Google and PayPal, two of the key players in an industry group known as FIDO, which stands for Fast IDentity Online Alliance. FIDO isn’t boosting any particular approach to identity recognition; mainly it plans to set industry standards. But it is promoting what’s known as two-step verification as a move in the right direction.
This is when you’d be identified by a combination of “something you know”—such as a password—with “something you have”—such as a token that plugs into your device’s USB port—or “something you are”—such as your fingerprint. This combo of a password and a device you carry around with you—Google security experts have suggested a log-in finger ring—would be a lot safer than a simple password, and would let you use an easy-to-remember password, since the account can’t be hacked without your ring or your fingerprint.
And once fingerprint sensors or face and voice recognition software become more common, it will be that much easier for passwords to simply fade away.
That feels inevitable to Michael Barrett, chief information-security officer of PayPal and president of FIDO. “Consumers want something that’s easy to use and secure,” he says. “Passwords are neither.”
A fingerprint scanner on your phone is only the beginning. There are a number of other inventive, and yes, even bizarre ideas for replacing passwords. Among them:
- Coming soon to a stomach near you: Let’s start strange. At a conference in late May, Regina Dugan, head of advanced research at Motorola, suggested that one day you’ll be able to take a pill every day that would verify your identity to all of your devices. The pill would have a tiny chip inside and when you swallow it, the acids in your stomach would power it up. That creates a signal in your body, which, in essence becomes the password. You could touch your phone or your laptop and be “authenticated in.” No, it’s not happening any day now, but the FDA has already approved its precursor—a pill that can send information to your doctor from inside your body. In other words, it’s a lot more plausible than it sounds.
- So, how about a tattoo that spells “password:” But that’s not all Dugan projected for the future. She also showed off an electronic tattoo. Motorola, now owned by Google, is working with a company named MC10, which has developed this “stretchable” tattoo with its own antenna and sensors embedded in it. It’s so thin, it can flex with your skin. And it would serve as your password, communicating with your devices and verifying that you are who you say you are.
- Now, what are all these keys for?: Back to the present. A Canadian company called PasswordBox is now offering a free app that remembers and automatically enters all your passwords across all your platforms. It signs you into websites, logs into apps, and enables you to securely share your digital keys with friends and loved ones—all through an app for your smartphone and a Chrome browser extension for your desktop. Its pitch is one-click login everywhere.
- Would my heart lie?: Another Canadian company called Bionym is building its business around the fact that heartbeats, like fingerprints, are unique. Its approach is to turn your heartbeat into a biometric pass code that’s embedded in a wrist band which, in turn, uses Bluetooth to let your machines know you’re the real deal.
Video bonus: Let’s go back to the future with John Chuang, a researcher at the UC Berkeley School of Information. He’s working on the idea of allowing people to verify their identities through their brain waves. Okay, at least hear him out.
Video bonus bonus: The Internet Password Minder is a stroke of…something. Even Ellen DeGeneres was impressed, in a funny way.
More from Smithsonian.com
How You Type Could Become Your New Password
November 30, 2012
A little refresher:
Back in late 2005, the guys running a small San Francisco startup named Odeo were feeling desperate. They had planned to make it big in the podcasting business, but Apple had just announced that iTunes would include a podcasting platform built into every iPod.
So the Odeo group started scrambling to come up with a new plan. One of the employees, a guy named Jack Dorsey, came up with the idea of a system where you could send a text message to a number and it would be delivered to all of your friends.
Someone came up with the code name twttr–a takeoff on Flickr–and when they looked up twitter in the dictionary and saw that it meant 1) A short burst of inconsequential information and 2) Chirps from birds, they agreed, Dorsey recalls, that the name “was just perfect.”
Such a tool
This is just to remind all of us that Twitter was born not as a grand vision, but more an act of desperation. And that it was originally meant as nothing more than a cool way to send reports of your status to all of your friends at once.
Which makes it all the more remarkable that these days Twitter is being hailed as everything from a barometer of the nation’s emotional health to a conduit for the flow of linguistic invention to a tool for urban planners to map travel routes.
Oh, and earlier this week, a young mother reportedly named her newborn daughter “Hashtag.”
There are those, of course, who think way too much is being made of Twitter’s capacity for capturing the zeitgeist. But there’s no question that it’s gaining status as an analytical tool. Here are just a few of the ways it’s being taken seriously:
1) It’s not the tweet, it’s emotion: Last month tech giant SGI rolled out something it calls the Global Twitter Heartbeat, a Big Data analysis of 10 percent of the roughly 500 million tweets tapped out every day.
The tool takes geotagged tweets over a period of time and converts them into a “heat map” designed to show the tone and intensity of what’s being said where. It’s first big effort was during Superstorm Sandy.
2) Pocket of politeness? Or pool of profanity?: The company Vertalab created its own Twitter heat map a few months ago, but that one focused on the use of two particular phrases on Twitter. While many weighed in with a conventional “Good morning,” a surprising number posted a two-word phrase rhyming with “duck flew.” .
True to form, the well-mannered tweets tended to bubble up from the South, particularly parts of Texas and Tennessee, while the cursing flowed freely around New York, Toronto and especially Los Angeles.
3) I hear ya, bruh: Researchers at Georgia Tech analyzed 30 million tweets sent around the U.S. between December 2009 and May 2011 and concluded that new words, at least on Twitter, tend to first pop up in cities with large African-American populations, then spread.
One example they gave was “bruh,” a Twitter version of “bro,” that first appeared in several cities in the U.S.’ Southeast, then leap-frogged to California.
4) The roads most traveled: Data-mapping expert Eric Fischer tracked millions of tweets from around the world and laid them over maps of highways to get a sense of how many people are heading where. He thinks urban planners could use this kind of data to fine-tune existing transportation systems and figure out where new routes are needed.
5) Exit polls are so last century: Go ahead and scoff, but some think Twitter analysis can even help predict an election. Barack Obama’s victory in the recent presidential race didn’t come as a big surprise to the Pew Research Center, which analyzed 2,500 online conversations in the two months leading up to the election. It found that a much higher percentage–58 percent–of the comments about Mitt Romney were negative, while 45 percent of the tweets about Barack Obama were harsh.
At the same time Twitter did its own analysis of which tweets by both campaigns provoked the strongest responses in which states. One key indicator: Obama had a high engagement level in the key swing state of Ohio–determined by retweets and favorites–while Romney had only a moderate engagement level there.
6) When military intelligence is not an oxymoron: Three U.S. Defense Department units are field-testing a software called the Dynamic Twitter Network Analysis (DTNA), to see how effective it is at gauging public opinion in political hot spots around the world. The software pulls in data from the public Twitter feed, then sorts it, live, by phrases, keywords or hashtags. The hope is that intelligence officers could use the software to understand people’s moods about a topic, or hopefully prevent or respond faster in any future U.S. embassy attacks.
7) I’m not a doctor, but I play one on Twitter: Researchers at the University of Pennsylvania were pleasantly surprised to see that people are using Twitter to share information on medical subjects that wouldn’t seem the stuff of tweets, such as cardiac arrest and CPR. Their analysis of a month of tweets found more than 15,000 messages that contained specific and useful information about cardiac arrest and resuscitation.
8) When short stories aren’t short enough: And finally, it is here at last, the first annual Twitter Fiction Festival. Since Wednesday two dozen authors from five continents have been posting their mini-stories in five different languages. The fare ranges from Iowa writer Jennifer Wilson posting photographs of gravestones, then writing “flash fiction” in response to epitaphs submitted by followers, to French fantasy novelist Fabrice Colin writing a serialized story of five strangers trapped on a bus. Stop by at the Fiction Festival website–it will be over before you know it.
Video bonus: Here’s another SGI heat map, this one tracking Obama and Romney-related tweets during election week.
More from Smithsonian.com
November 27, 2012
Black Thriday is over. So is Small Business Saturday and Cyber Monday. Today, in case you didn’t know, is either Green Tuesday or Giving Tuesday, depending on whether you feel like eco-shopping or giving to charity.
Not sure what tomorrow may bring (How about Weird Relative Gift Wednesday?), but I suppose shopping does feel less chaotic if someone’s organizing it into theme days, although that doesn’t always stop it from devolving into a contact sport.
Can you imagine American shoppers embracing something like iButterfly, a mobile app popular in Asia where customers earn coupons by tracking down virtual butterflies with their smartphones? Me neither.
In the U.S., it’s about cutting to the chase and here the chase is after the sweetest deals, pure and simple, without having to bother with running after faux flying insects. And retailers have ratcheted up the competition, using the latest tracking technology to closely monitor their competitors’ pricing decisions and undercut them, in close to real-time, on their own websites. When Best Buy, for instance, published advertising saying it would be selling a $1,500 Nikon camera for $1,000, Amazon responded on Thanksgiving morning by cutting its price for the same camera to $997.
To know you is to lure you
No question that the big hook remains big bargains. But a lot of companies are also getting much more aggressive about mining data to tap into the power of personalization. The more they know about you and your tastes and habits and what you say on Facebook, the more they can press your buy buttons–but in a way that feels like they’re doing it all for you.
Now grocery stores like Safeway and Kroger have even started to customize prices in offers to loyalty cardholders. As Stephanie Clifford noted in the New York Times:
“Hoping to improve razor-thin profit margins, they are creating specific offers and prices, based on shoppers’ behaviors, that could encourage them to spend more: a bigger box of Tide and bologna if the retailer’s data suggests a shopper has a large family, for example (and expensive bologna if the data indicates the shopper is not greatly price-conscious).”
And RetailMeNot, the most popular coupon site in the U.S., has just launched an app that steers you to coupons you’re more likely to use based on your Likes and other personal info gleaned from Facebook.
But when does solicitousness turn creepy? Is it when you receive a pitch in your email for an outfit you pinned on Pinterest? Or when you start getting offered bargains from stores you happen to pass on the way to work every day?
If you believe a recent survey by Accenture Interactive, a clear majority–61 percent–of online shoppers in the U.S. and the U.K. are willing to give up some privacy if it means they can receive personalized offers from retailers.
And more than 50 percent of those surveyed in the U.S. said they’re comfortable with the idea of their favorite retailers tracking their personal data in order to fine tune recommendations for future purchases.
But only so comfortable. Almost 90 percent of the respondents said that’s entirely dependent on whether retailers offer them choices on how their personal info can be used.
As Kurt Kendall, a retail consultant, put it in a recent interview with Cox Newspapers: “People do not want to feel like they’re being stalked.”
I’ve got my fake eye on you
How about being watched? The obsession with gathering intelligence about customer behavior has reached the point where an Italian company is selling mannequins equipped with cameras to watch shoppers. This model, called the EyeSee, is being sold by Milan-based Almax for more than $5,000.
That’s a lot of money for a pretend person. But this one has a camera embedded in one eye that feeds data into facial-recognition software which logs the age, gender and race of passers-by. It’s all about collecting data–no video is actually stored.
Almax won’t reveal which of its clients have purchased EyeSee mannequins, but it has said that one added a children’s line of clothing when the camera observed that kids made up more than half its mid-afternoon traffic. Another, according to Almax, discovered that a third of its visitors using one of its doors after 4 p.m. were Asian, prompting it to place Chinese-speaking staff by that entrance.
But wait, there’s more. Almax is developing a model that will recognize words well enough that stores will be able to find out what customers are saying about the mannequin’s outfit–again without recording a thing.
The shipping news
Here are more examples of how companies are using technology to build relationships with customers.
- Or simply “Clothes That Don’t Make Me Look Fat”: For those who know what they like in fashion, Shop It to Me has just launched a site called Shop It to Me Threads that allows you to create a customized page that’s updated daily with the latest news and deals on your favorite fashion trends, designers, types of items, or combination of elements, such as “Michael Kors Bags and Shoes under $250″ or “Pencil skirts under $100.”
- Pickie picky: E-commerce start-up Pickie has come out with an iPad app that builds a personalized shopping catalog for you, based on your preferences expressed on Facebook, along with suggestions from your friends. And you’re able to order items directly from your customized Pickie site.
- Do it for the children: To counter the trend called “showrooming,” where people check out products in a store and then go home and buy it from another company online, Target is encouraging shoppers to go online while they’re in its stores. During the holidays, the retailer is featuring 20 hot toys at the front of its stores next to signs with QR codes. Shoppers with smart phones can scan the codes, buy a toy and have it shipped free.
- What about Pop Tarts and headphones?: Amazon, through its subsidiary Quidsi, is sharpening its aim at moms who shop online. Last month it launched another narrowly targeted site called AfterSchool.com. It lists more than 70,000 of the sort of things kids need after school, from ballet shoes and shin guards to basketballs and jewelry kits.
- And if you’re really loyal, a greeter washes your car: Earlier this month Walmart, through its Silicon Valley operation @WalmartLabs, rolled out Goodies, a food subscription service. For $7 a month, people who sign up will receive a box of gourmet snacks, such as Dang Toasted Coconut Chips and a Nutella & Go snack pack. And if they’re active on the Goodies site by rating products and writing reviews, they can earn enough loyalty points to start getting their monthly goodies for free.
Video bonus: Based on this video from Comiket, the huge comic book convention held in Tokyo, the Japanese and Americans have very different styles when it comes to the surging crowd thing.
More from Smithsonian.com
November 9, 2012
This is what election night is like in America these days:
I had gathered with about dozen other people, ostensibly to watch the results on TV. But the TV received, at best, divided attention.
To my left, my wife Carol had fired up her laptop and was foraging for results on websites that might have vote totals more current than what was on the big screen. To her left, another woman was zeroed in on her smart phone and to my right, two more guests were doing the same. So was I, for that matter. I kept one eye on the TV so I didn’t miss any states changing color, but my good eye was focused on my smart phone, where I was following the running commentary of Facebook friends.
Of the people in the room, at least half were furiously working another screen.
And then, when NBC called the election for Barack Obama, our hostess jumped up and, with her smart phone, snapped a picture of the announcement on the TV screen, closing, for one fleeting moment, the screenfest loop.
Earlier that same day, appropriately, the Norwegian company never.no launched an interactive content tool called Sync. It’s designed to give advertisers the opportunity to jump on to the second screen so a commercial gets the attention for which the sponsor has paid. But we’re not talking about just showing the same ad at the same time on a smaller screen. That would be both lame and annoying.
No, Sync is meant to actually put an ad in play on the screen where the action is. You’d be encouraged to interact with it–answering poll questions, getting more info about a product, maybe even sharing a clip about it on Facebook and Twitter. And as this approach gets more sophisticated, the thinking goes, it will become possible to flip things around so that the audience can influence an ad in real time, perhaps by selecting an ending from several different choices.
For advertisers this would be a beautiful thing–genuine viewer engagement in an experience that makes an ad personal and extends its life beyond its 30 seconds on screen. All while tracking the behavior of all those people interacting with it.
Screen on me
Other companies have also been trying to master the two-screen shuffle, including Shazam, the outfit best known for creating the mobile app that can tell you the name of a song once it hears the music. Starting with the Super Bowl last February, when it worked with more than half of the event’s advertisers to steer owners of its app to bonus content, Shazam has been refining the process of using mobile phones to connect viewers in more personal ways to TV programs and advertisers.
It still follows its original concept of recognizing sounds or music to identify a show or sponsor, but now it takes the next step of actually providing opportunities to bond with a product.
The latest example rolled out in Ireland a few days ago, an ad for Volvo. Anyone with the Shazam app on their phone–and there reportedly are now more than 250 million people around the world who have it–can “tag” the Volvo ad when it comes on TV and that, among other extras, allows them to then sign up for a free test drive and get a chance to win an iPad mini.
Take this personally
Okay, but how many of us really want to engage with a commercial? Don’t we do just about anything to avoid watching them? People in the multi-screen business acknowledge this. They know people tend to resent the intrusion of ads into the personal space of their phones and that many would much rather play Words With Friends during commercials than get all chummy with a bathroom cleaner.
And yet while recent research found that at least three out of four TV viewers say they use some other device while watching, a nice chunk of them–more than a third–say they’ve used their cell phone or digital tablet to browse for products spotted in a show or ad.
So the inclination is there. The key for advertisers is learning to create true value for viewers in the experience they provide on the small screens, a real reason to interact, not just some shrunken message of what they put on the TV screen.
Which brings me back to the election. There’s already talk that four years from now, political advertising will need to move into the multi-screen world of the 21st century. It will need to evolve beyond the thinking that volume is everything, that the days are over when the winner invariably was the side that could hammer home its message most often.
A case in point: An analysis of Super PAC spending published this week by the Sunlight Foundation found that American Crossroads, which spent more than $100 million on campaign advertising this year, had a success rate of just 1.29 percent.
Here are more recent developments in efforts to reach people on multiple screens:
- Life imitates TV: NBC will begin using a social TV app called Zeebox, which not only allows viewers to converse in real time with friends watching the same show, but also now will provide them with info on how they can purchase items in shows, particularly clothing and kitchen products.
- When you wish you were a star: A live ad for the recent launch in Great Britain of the popular Xbox video game Halo 4 featured a “roll call of honor,” a display of the names and pictures of randomly selected gaming fans who opted in via Facebook. The ad also showed, in real time, the number of people playing Halo 4 on Xbox Live.
- You make the call…in 140 characters or less: Also in the U.K., a recent campaign for Mercedes-Benz allowed viewers to vote on Twitter to determine how an ad featuring a chase scene should end.
- Will only redheads see ads for ginger snaps?: Earlier this fall Allstate worked with DirecTV and the Dish Network to target the audience so that only renters saw an ad for renter’s insurance.
Video bonus: Here’s a taste of the Mercedes-Benz ad that viewers controlled through Twitter.
More from Smithsonian.com