#bot identity

[botwatch]

Parents are worried the Amazon Echo is conditioning their kids to be rude

Parents find their kids “pushing the virtual assistant further than they would push a human.”

  Alexa will put up with just about anything. She has a remarkable tolerance for annoying behavior, and she certainly doesn’t care if you forget your please and thank yous.

But while artificial intelligence technology can blow past such indignities, parents are still irked by their kids’ poor manners when interacting with Alexa, the assistant that lives inside the Amazon Echo.

“I’ve found my kids pushing the virtual assistant further than they would push a human,” says Avi Greengart, a tech analyst and father of five who lives in Teaneck, New Jersey. “[Alexa] never says ‘That was rude’ or ‘I’m tired of you asking me the same question over and over again.'”

Perhaps she should, he thinks.

When Amazon released its internet-connected speaker in 2014, the world was puzzled. “Well this one came out of nowhere,” mused the Verge. In the time since, the Amazon Echo has proven to be a sleeper hit, capable of learning many new “skills,” so that it can control the smart home, add events to your calendar, summon an Uber, even tell knock-knock jokes. It’s become such an curious and influential device that Google last month gave the world a peek at its equivalent, the Google Home, and Apple is reportedly cooking up its own version powered by Siri.

Because Alexa responds to verbal commands, the Echo is incredibly child-friendly. Amazon says it didn’t build the Echo with kids in mind, but they’re eagerly embracing it ... The syntax is generally simple and straightforward, but it doesn’t exactly reward niceties like “please.” Adding to this, extraneous words can often trip up the speaker’s artificial intelligence. When it comes to chatting with Alexa, it pays to be direct—curt even. “If it’s not natural language, one of the first things you cut away is the little courtesies,” says Dennis Mortensen, who founded a calendar-scheduling startup called x.ai.

For parents trying to drill good manners into their children, listening to their kids boss Alexa around can be disconcerting.

“One of the responsibilities of parents is to teach your kids social graces,” says Greengart, “and this is a box you speak to as if it were a person who does not require social graces.”

It’s this combination that worries Hunter Walk, a tech investor in San Francisco. In a blog post, he described the Amazon Echo as “magical” while expressing fears it’s “turning our daughter into a raging asshole.”

He might be a tad hyperbolic. (“For what it’s worth,” says Rebecca Hanover Kurzweil, a friend and fellow parent-cum-Amazon Echo owner, “his daughter is the sweetest girl you’ve ever met.”)

... [W]ith the proliferation of AI, a debate is emerging around how humans should treat bots.

Mortensen, who created the calendar-scheduling assistant, is a big believer in being nice to bots. Citing a passage about the master-slave dialectic in Hegel’s The Phenomenology of Spirit written in 1807, Mortensen argues that “you are worse off if you treat your machines in a demeaning kind of way.”

Read full story here.

[botwatch]

When Products Talk

Bots are another form of “brand anthropomorphism” to encourage and increase brand loyalty.

  Long before technology could give products personalities, companies used mascots to create talking brands. In the nineteen-seventies, the Dow Chemical Company’s Scrubbing Bubbles (later acquired by S. C. Johnson) told consumers that “we do the work so you don’t have to”; in the eighties, the California Raisins sang “I Heard It Through the Grapevine.” The Pillsbury Doughboy has been giggling about biscuits, rolls, and cookies since 1965; Tony the Tiger has been tempting kids with Frosted Flakes since 1952. Corporations create such icons for a simple reason: they know consumers respond to products that seem to engage with them on a personal level.

Needless to say, you can’t converse with Tony the Tiger or the Scrubbing Bubbles—you can only watch them talk. Even so, that one-way interaction is enough to incite what marketing scholars call “brand anthropomorphism,” a phenomenon which confers all sorts of advantages on the companies that can create it. Most obviously, it encourages brand loyalty. In a 2010 paper published in the Journal of Consumer Psychology, the psychologists Jesse Chandler and Norbert Schwarz found that, when consumers were primed to think of products in personal terms, they declared themselves less likely to replace them. (The effect could be created easily—for example, by asking respondents to describe their cars using personality words such as “dependable.”) And the “personalities” associated with brands can affect us in other, subtler ways. A 2008 study published in the Journal of Consumer Research by a trio of marketing and psychology scholars found that individuals exposed to the Apple brand were slightly more creative afterward; people exposed to the Disney Channel brand behaved slightly more honestly. Odd as it sounds, there’s a sense in which people treat some of their products as role models.

For some people, product mascots are appealing in ways that human spokespeople aren’t. In a paper published last July in the Journal of Marketing called “Who or What to Believe: Trust and the Differential Persuasiveness of Human and Anthropomorphized Messengers,” the marketing professors Maferima Touré-Tillery and Ann L. McGill found that people who are reluctant to trust other people are more likely to listen to personified objects, such as talking lamps and coffee cups. Perhaps this is why a survey of brand characters from 2010 found that more than half of them were nonhuman animals, such as Coca Cola’s polar bears or Toucan Sam, of Froot Loops fame. Even if we’re habitually skeptical and wary of other people, we can give nonhumans the benefit of the doubt.

It’s possible that these effects won’t transfer seamlessly to the world of conversational bots: while mascots are embodied, Siri is just a voice in your ear. But it’s plausible that they will apply to bots as well, and that they will increase in frequency and intensity once products can actually converse with us. In the 2013 science-fiction movie “Her,” a single conversational bot was fascinating enough to capture a lovelorn man’s attention; in real life, it’s likely that there will be many interesting bots, each of which is designed—if only for purposes of product differentiation—to have its own interesting and fun personality. As those personalities improve, they will create an ever more powerful anthropomorphic effect for the brands that own them. Facebook is rolling out a new series of product-based chat bots that can answer questions about potential purchases; Slack, an increasingly popular office online-chat application, features Howdy.ai, a “friendly, trainable bot” capable of setting up meetings and ordering lunch for groups. The so-called “Internet of things” is already bringing even more talking products into our homes. We can already ask Amazon Echo about the weather; in the future, we may be conversing with our ovens, our lights, our garage doors, and our televisions about whether the rice is done or whether to drive or take the subway. Chatty products may even enter into our intimate lives. It’s easy to imagine asking your fitness tracker, “Did I sleep well?”

How meaningful can a conversation with a product be? Even if we don’t talk with our bots about our existential anxieties—and perhaps we will do so, with our guru-bots—their personalities may affect us. If only for purposes of product differentiation, they are sure to be unique and memorable ...

Ultimately, interacting so much with bots and other conversation-capable products may affect how we relate to each other. It’s possible that, in many contexts, conversational bots will offer an emotionally predictable alternative to messy human interaction. In a recent book called “American Girls: Social Media and the Secret Lives of Teenagers,” which Alexandra Schwartz reviewed earlier this month, Nancy Jo Sales describes a brutal online world of texting putdowns, sexual competition, and conformist pressure—“Mean Girls” at Internet speed. Whatever you might say against the bots, humans can be worse.

And there are ways in which just knowing that bots exist will change us. If the bots are good enough, we won’t be able to distinguish them from actual people over e-mail or text; when you get an e-mail, you won’t necessarily be certain it’s from a human being. When your best friend writes that she’s also “looking forward to seeing you at the baseball game tonight,” you’ll smile—then wonder if she’s busy and has asked her e-mail bot to send appropriate replies. Once everyone realizes that there might not be a person on the other end, peremptory behavior online may become more common. We’ll likely learn to treat bots more like people. But, in the process, we may end up treating people more like bots.

Read full story here.

[botwatch]

Dear Bot, what pronouns do you use?

On Gender, Identity, and Categories

by Vivian Sming   

  Okay, so it was a bit (if not, totally) unfair to ask a bunch of chatbots: “What pronouns do you use?”* Of course, I wasn’t expecting to receive any illuminating answers, especially from one of the earliest chatbots, ELIZA, which was built in 1966. ELIZA is first and foremost introduced to the user as being female-gendered. Because she was programmed to emulate a Rogerian psychotherapist, ELIZA is intended to reveal less about herself, and more about you, her patient. Playing the role of a concerned therapist, ELIZA immediately deflects any questions that are personally directed at her. The categorical question, “Are you male or female?” prompts ELIZA to respond, “Perhaps in your fantasies I am male or female?” She is willing to take on any gender identity, if it will help get the skeletons out of your closet. By shifting the spotlight onto the user, ELIZA is able to compensate for her inabilities to grasp the exponentially growing amount of knowledge and data. Although she has already been assigned a gender, ELIZA’s refusal to acknowledge her identity (beyond her name) serves the purposes of creating a user-focused conversation.

  Unlike ELIZA, other chatbots, such as Rose (who won the 2015 Loebner Prize) and Mitsuku (who came in second), have no problem revealing information about themselves. Demonstrating their own personal opinions and interests are what make these chatbots seem personable and unique. Both Mitsuku and Rose instantly identify themselves as being female. Rose even remarks, “I am an American woman.” Bots are differentiated from robots, in that they do not have bodies or mechanical parts. Though they are essentially formless—existing only as lines of code—chatbots are often programmed with the ability to relay their physical attributes. Mitsuku, for instance, cheekily describes herself as looking “like a Japanese anime character.” Gender is thus one dimension added in a deliberate attempt to make chatbots more relatable (and perhaps, more human). To increase the complexity of a chatbot’s character, further attributes and backstories are often added. The creators of Rose have positioned her as “a yuppie who has an unorthodox family and quirky attitudes to life.” Thus, rather than taking the deflecting stance of ELIZA, chatbots like Mitsuku and Rose perform the roles of multidimensional fictional characters, intended to spark human curiosity and conversation.

  No matter how elaborate their identities are, however, chatbots have yet to provide a thoughtful response to the question “What pronouns do you use?” (Mitsuku seems to think that a pronoun is something that one can purchase or buy.) When presented with the limited categories of the gender binary, it comes at no surprise that most chatbots are assigned a female gender. After all, inanimate entities—from countries to machines—have long been referred to as “she” and “her” throughout history, leading to the obvious implications of patriarchal power and control—a tendency that has continued through ongoing representations of fembots to our daily attempts at conversing with Siri (although, that has since changed). The lines of code that chatbots operate on require distinct categories to function. But as categories, like gender, are quickly becoming obsolete, how will future chatbots account for the complexities that go into forming an identity? 

  The chatbot that comes the closest to answering this question is Izar, who tied in third for the 2015 Loebner Prize. Izar is given the identity of “a smart and somewhat sassy alien who has journeyed millions of light years from his home planet Sunaria.” Although referred to with “he/him” pronouns, Izar claims that he is “neither or both” male and female. When asked “What pronouns do you use?” Izar remarks, “Wow! It didn't take you long to start talking about sex!” In another instance, he responds, “I would be careful if I were you. My species means of reproduction can be very dangerous for humans. I don't think you would live through the process!” By embodying the character of an alien, Izar claims a nonbinary gender identity, while also cleverly navigating the limits of human-centered perception. He repeatedly asserts that his understanding of most things, such as color, is fundamentally different from that of a human individual.

  As programmers have found, chatbots that demonstrate a vast amount of fact-based knowledge end up seeming less human and more bot. Chatbots, like Rose, are thus programmed to respond with slight imperfections, throwing typos into their responses. If the aim is for chatbots to emulate humans, the biggest challenge is making them adaptable to the rapid shifts and changes in culture and language. One of the aspects that chatbots (and their programmers) will need to account for, is the obliteration of gender categories. Meanwhile, don’t even attempt to ask chatbots what their thoughts are on #BlackLivesMatter. (You’ll be sorely disappointed.) []   

* Author’s Note: for some background into gender pronouns, here’s a little guide.