The Most Human Human (3 page)

In the months before the test, I did everything I could to prepare, researching and talking with experts in various areas that related back to the central questions of (a) how I could give the “most human” performance possible in Brighton, and (b) what, in fact, it means to be human. I interviewed linguists, information theorists, psychologists, lawyers, and philosophers, among others; these conversations provided both practical advice for the competition and opportunities to look at how the Turing test (with its concomitant questions of humanhood) affects and is affected by such far-flung fields as work, school, chess, dating, video games, psychiatry, and the law.

The final test, for me, was to give the most uniquely human performance I could in Brighton, to attempt a successful defense against the machines passing the test, and to take a run at bringing home the coveted, if bizarre, Most Human Human prize—but the ultimate question, of course, became what it
means
to be human: what the Turing test can teach us about ourselves.

National Public Radio’s
Morning Edition
recently reported the story of a man named Steve Royster. Growing up, Royster assumed he had an incredibly unusual and distinctive voice. As he explains, “Everyone always knew when I was calling just by the sound of my voice, while I had no earthly
idea
who was on the phone when
they
called.” It would take him until his late twenties before he fully grasped—to his amazement—that other people could discern most
everyone’s
identity by voice. How on earth could they do that? As it turns out, there
is
something unusual about Royster, but not about his voice: about his brain. Royster has a rare condition known as “phonagnosia,” or “voice blindness.” Even when Royster’s own mother calls him, he simply goes politely along with the flow of the conversation, unaware that “this strange woman who has called me is, in fact, the one that gave birth to me.” As reporter Alix Spiegel puts it, “Phonagnosics can tell from the sound of your voice if you’re male or female, old or young, sarcastic, upset, happy. They just have no blooming idea who you are.”

This all puts Royster, of course, in an awfully strange position.

It happens to be the same position everyone is in on the Internet.

On September 16, 2008, a twenty-year-old college student named David Kernell attempted to log in to vice-presidential candidate Sarah Palin’s personal Yahoo! email account. He didn’t have a clue what her
password might be. Guessing seemed futile; instead, it occurred to him to try to
change
it—and so he clicked on the “I forgot my password” option available to assist absentminded users. Before Yahoo! will let a user change an account password, it asks the user to answer several “authentication” questions—things like date of birth and zip code—in order to “Verify Your Identity.” Kernell found the information on Wikipedia, he said, in approximately “15 seconds.” Stunned, Kernell “changed the password to ‘popcorn’ and took a cold shower.” Now he faces up to twenty years in prison.

In the world of machines, we authenticate on
content:
password, PIN, last four digits of your Social Security number, your mother’s maiden name. But in the human world, we authenticate on
form:
face, vocal timbre, handwriting, signature.

And, crucially, verbal style.

One of my friends emailed me recently: “I’m trying to rent a place in another city by email, and I don’t want the fellow I’ve been communicating with to think I’m scamming him (or, am a flake), so, I’ve been hyperaware of sounding ‘human’ and ‘real’ and basically ‘nonanonymous’ in my emails. A weird thing. Do you know what I mean?” I do; it’s
that
email’s idiosyncrasies of style—the anachronistic “fellow,” the compound, unhyphenated “hyperaware” and “nonanonymous”—that prove it’s really
him
.

This kind of thing—behavior that seems “so you”—might always have been, say, charming or winning (at least to those who like you). Now it’s something else too, our words increasingly dissociated from us in the era of the Internet: part of online
security
.
¹

Antarctic penguins detect the precise call of their chicks among the 150,000 families at the nesting site. “Bless Babel,” fiction writer Donald Barthelme says. It’s true: ironing out our idiosyncrasies in verbal style would not only be bad for literature; it would be bad for
safety
. Here as elsewhere, maybe that slight machine-exerted pressure to actively assert our humanity with each other ends up being a good thing.

One of my old college friends, Emily, came into town recently, and stopped downtown on her way from the airport to have lunch with a mutual friend of ours and his co-worker—who happened also to be my girlfriend, Sarah. When Emily and I met up later that day for dinner, I remarked on how funny it was that she’d already met Sarah before I’d had any chance to introduce them. I remember saying something to the effect of, “It’s cool that you guys got to know each other a little bit.” “Well, I wouldn’t say that I got to
know
her, per se,” Emily replied. “More like, ‘saw what she’s like’ or something like that. ‘Saw her in action.’ ”

And that’s when the distinction hit me—

Having a
sense
of a person—their disposition, character, “way of being in the world”—and knowing
about
them—where they grew up, how many siblings they have, what they majored in, where they work—are two rather different things. Just like security, so does intimacy have both form and content.

“Speed dating” is a kind of fast-paced, highly structured round-robin-style social mixing event that emerged in Beverly Hills in the late 1990s. Each participant has a series of seven-minute conversations, and at the end they mark down on a card which people
they’d be interested in meeting again; if there are any mutual matches, the organizers get in touch with the relevant contact information. Though it’s entered into popular parlance, “SpeedDating” (“or any confusingly similar term”) is technically a registered trademark, held by, of all groups, the Jewish organization Aish HaTorah: its inventor, Yaacov Deyo, is a rabbi.

One of my earliest thoughts about the Turing test was that it’s a kind of speed date: you have five minutes to show another person who you are, to come across as a real, living, breathing, unique and distinct, nonanonymous human being. It’s a tall order. And the stakes in both cases are pretty high.

A friend of mine recently went to a speed-dating event in New York City. “Well, it was the oddest thing,” he said. “I kept wanting just to, like, banter, you know? To see if there was any chemistry. But all the women just kind of stuck to this script—where are you from, what do you do—like they were getting your stats, sizing you up. But I don’t care about any of that stuff. So after a while I just started giving fake answers, just making stuff up, like. Just to keep it interesting, you know?”

The strangeness he experienced, and the kinds of “bullet points” that speed dating can frequently devolve into, are so well-known as to have been lampooned by
Sex and the City:

The delivery certainly isn’t.

People with elaborate checklists of qualities their ideal mate must have frequently put entirely the wrong types of things. This height.
This salary. This profession. I’ve seen many a friend wind up, seemingly unsuspecting, with a jerk who nevertheless perfectly matched their description.

Fed up with the “Dwight Owens”–style, salvo-of-bullet-points approach that kept recurring in early speed-dating events, Yaacov Deyo decided on a simple, blunt solution: to make talking about your job
forbidden
. People fell back on talking about where they lived or where they were from. So he made that forbidden too. He seems charmed and maybe even a little smug enacting the ensuing panic, then breakthrough: “Omigosh, like, what
do
I talk about?” He laughs. “I can’t talk about what I do for a living, or where I live, and … wow! All of a sudden I have to describe me.” Or: all of a sudden I have to
be
me, to
act like
myself instead of
describing
myself.

The first Loebner Prize competition was held on November 8, 1991, at the Boston Computer Museum. In its first few years, the Loebner Prize gave each program and human confederate a “topic” as a means of limiting the conversation. In some ways, that first contest was a battle between style and content. One of the confederates was Shakespeare expert Cynthia Clay, who was, famously, deemed a computer by three different judges. The consensus seemed to be “No one knows that much about Shakespeare.” (For this reason, Clay took her misclassifications as a compliment.)

The program behind the topic of “whimsical conversation”—PC Therapist III by Joseph Weintraub—won the Loebner Prize that year, earning the very first Most Human Computer award. As the program practiced it, “whimsical conversation” was not a
topic
at all, but a
manner
—a crucial distinction. Weintraub was able to play off of a looseness in the contest rules, the natural behavior patterns of chatbots, and people’s stereotypes about computers and humans all at once. Strategically, it was brilliant. Clay and her fellow confederates,
by dutifully and scrupulously providing information to the questions asked, demonstrated their knowledge and understanding—but sometimes in a factual, encyclopedic way commonly associated with computer systems.

The interlocutor at the “whimsical conversation” terminal stood in stark contrast:

“I really thought [PC Therapist] was human because it … was weird and funny, in a normal sort of way,” said one of the judges, a Harvard graduate student. Weintraub’s program, shifting topics wildly and spouting non sequiturs and canned one-liners, came off as zany, a jokester, a much more “human” personality type. At least—before I learned how easy it was to mimic—I used to think so.

Looking around on the Internet for information about the Most Human Computer of 2005 and 2006, Rollo Carpenter’s web-based Cleverbot, I found something very curious. There seems to be a considerable amount of speculation that the site may be, in effect, a hoax.