Inside My Days as a Content Bot
"Hearst Magazines and Yahoo may earn commission or revenue on some items through these links."
In the summer of 2007, I was freshly out of college and searching for jobs with increasing desperation. I had worked unpaid internships at a newspaper and a literary journal; I’d written music reviews. None of this, it was becoming apparent, screamed “employable.” Who would pay me to do what I wanted to do, which was write? When I happened across a Craigslist ad that said, “Creative writer wanted,” I responded. My recent job inquiries had been as effective as dropping bottled messages into the Pacific. When a response came—“Can you come into our Menlo Park offices this week for an interview?”—I replied immediately. Yes, I could.
Did I wear a blazer? Did I take Caltrain? How did I get around before Uber or GPS? I remember being nervous, and I remember wanting to be liked. I don’t remember what was asked of me, or how the job was described. The man interviewing me was older than I was then, younger than I am now. The job was “a lot like working for the CIA,” he told me. “Your best work, you can’t show anyone.”
These were the early years of Facebook, of shareable iTunes libraries—of downloading music illegally (or legally, but let’s be real). I had a phone that flipped open and didn’t take pictures. Menlo Park wasn’t on the map. It would be years before Mark Zuckerberg was named Time’s person of the year, and even more years before he came to be viewed as a creepy robot. Silicon Valley had not yet accrued its sinister air. Twitter had been founded recently, in March 2006, but it had not yet attracted media personalities; it wasn’t where anyone went for news. It hadn’t yet been renamed X by a billionaire, who wasn’t yet a billionaire. There were fewer billionaires back then. But Google existed, and so did its pages of search results. A search for a semi-prominent individual yielded information about them.
These search results would be the main focus of my new job. I was to be a protector of online reputations.
None of our clients had been convicted of any crimes, our higher-ups assured us. If ever they were, they would be dropped as clients. But everyone deserved to put their best foot forward—didn’t they? We would be helping them to do so. Each of our clients had hired us because they disliked their search results. For one client, the problem was that a New York Times Vows article about his marriage to his ex-wife was the very first link you saw when you searched his name. We couldn’t eliminate these offending items from the Internet. But we could try to bump them off Google’s top spot—or, even better, off the first page entirely, into the realm of the never-seen. All we had to do was create new material—what we called “pink sites.” This was what put the “creative” in “creative writer.”
For each client, we’d receive lists of their hobbies, interests, and positive attributes. These were the qualities we’d highlight. One client, a businessman, was an avid sailor in addition to a cheating spouse. We chose to emphasize the former. I created several websites about boating, mentioning his name on each. Then I would go about trying to promote each website using early social-media platforms: Reddit, Maple, StumbleUpon, Xing. “Check out this website on boating!” I might have posted on Twitter, from one of my dozens of made-up profiles. I didn’t bother to make my alter egos particularly realistic (one was “Salvador Whippet”). We were urged to make our websites as believable as we could, so they’d gain organic traffic and appear higher in search results—edging out less desirable sites.
Another method of reputation management was putting up websites about fictional people who happened to share our clients’ names. One of my coworkers maintained a stream-of-consciousness skateboarding blog, ostensibly penned by someone with the same name as a client but with a different age and a different personality. (“That ramp was hella sick but I was ready for it.”) We maintained spreadsheets of our pink sites, called pink profiles. I generated long lists of negative domain names that our company would purchase, so that our clients’ detractors wouldn’t be able to snap up the most desirable ones (soandsosucks.com, soandsoistheworst.com, ihatesoandso.com). These methods, combined, were called the “secret sauce.” The “secret” was heavily stressed. We called each client by a code name corresponding to their initials: Jennifer Aniston, Mark Wahlberg, Spider Man.
I was paid eight cents per word, so writing that didn’t require heavy research was ideal. For one site, I wrote short stories and poems from the perspectives of twelve-year-old girls. I mentioned the client’s name on one of the site’s pages, but beyond that, the site needed content. What was easiest for me to generate was subpar fiction and poetry. In one story, a girl living on Jupiter wakes up to find that her parents have disappeared. Plaintively, she wonders: “WHERE ARE MY PARENTS/ALL OF JUPITER?” My boss encouraged experimentation because we never knew what sites would catch on. If I succeeded in attracting real preteen girls to my invented site, their traffic might cause Google to rank the site more highly, deposing our client’s negative search results.
The office building was small and nondescript. I remember gray, even as I can’t picture the space anymore. We brought our own laptops to work. We chatted all day long on Instant Messenger—trying to make one another laugh, often succeeding. My coworkers were a ragtag group ranging in age and racial background. If not for the job, I might have never met them. They were excellent writers—strange and funny, united in our bizarre labor. It occurs to me now that we were bots before bots were bots. Did we call it “content” back then? I don’t think the noun had yet acquired its negative connotation, to mean meaninglessness. But none of us were under the illusion that what we were creating was lasting. It wasn’t the ideal job for any of us. It was just the job we had.
Every month, I aimed to write 40,000 words. It was a novel every one and a half months, though that meant I wasn’t working on my own hypothetical novel. My wrists hurt; I got carpal tunnel. What I wrote was fictional, but it wasn’t the fiction that I hoped to write. Within two months, I was desperately applying to MFA programs.
On weekends, I took my laptop to coffee shops. In those days, San Francisco cafés weren’t white-walled or third-wave. They served dark roasts and had sticky floors and faux-leather chairs that hissed. What I hoped to write was something of my own, something meaningful to me. But I never wrote anything good. It would take time to become a better writer—time I hadn’t lived yet.
My job was online, but my real life was lived in person. We didn’t need the acronym IRL back then. It was a given. Life was lived in real life—where else would it be lived? Facebook wasn’t widespread; Instagram didn’t exist.
I was living in a Victorian flat in San Francisco’s Lower Haight. The rent was $3,280 split four ways. My roommates were also new college graduates. We hosted dinners and parties—homemade pasta and pickles—and pooled our money for communal bourbon. The frequency of our gatherings feels like a relic of the past, too, though it’s hard to say if it’s that times changed or we did, settling into families of our own. When I told new friends what my job was, their eyes widened. You should write about that, they always said.
I remember one visitor, a friend of a friend. She was slightly older than we were—more cultured. She was a chef at a prestigious restaurant where I could not afford to eat. When I told her about my job, she could not conceal her disgust. She asked, “How do you sleep at night?” I remember being speechless at her judgment. I slept at night because I was tired, and this was the only job I had gotten.
amazon.com
$26.10
Seventeen years later, my experience has been reduced into a tellable anecdote. Most of my memories have fallen away, but I remember that moment, feeling judged by her. From her privileged position, she could judge me for my job. I can judge myself for it, now. My coworkers and I were doing what the Russian and Chinese hackers did to the American people before the 2016 election, albeit less effectively. None of our clients have been convicted of crimes or accused of sexual assault went my superiors’ constant refrain. But they weren’t exactly upstanding people.
In the years after writing content for the client we called Ed Bradley, I would happen across his name in the news: corruption, bribery, a sex scandal. His current Wikipedia page, which calls him a “disgraced former American lobbyist and businessman,” doesn’t even mention whatever he was up to in the years that I was generating websites on his behalf. Those particular misdemeanors seem to have vanished with time, replaced by more up-to-date offenses. It doesn’t appear that he changed his ways. He seems to be doing fine—still wealthy, largely unscathed. He pled guilty to a felony that carried a prison sentence of up to five years but was pardoned by Donald Trump.
I did become a writer. I write novels now, as I hoped to then. I have privilege enough not to have to work for eight cents a word. My writing can take more time. I changed, and the world did, too.
The Internet has become less personal. It feels endless—and not in a good way. Our online experiences have become angrier, more algorithm driven. In 2007, our lives were only beginning to have an online footprint. Since then, our existences have steadily moved in the virtual direction: Thanks to social media, smartphones, and omnipresent cameras, our online identities have become nearly synonymous with our identities themselves. The divide between public and private has eroded. Online reputations matter more than they ever have, and the methods for controlling narratives have become more powerful, too.
I’m not proud that I contributed to the Internet’s general bullshit quotient. But my coworkers and I were a dozen writers in a gray room in Menlo Park, and there was a limit to how much damage we could cause. With AI now available to generate writing, we are on the verge of infinite damage. In a John Henry-style contest to see who can write more preteen stories about Jupiter, AI would leave me dead at my keyboard. And AI can do much more than write stories about Jupiter: it can bend reality entirely. What I did back then makes for a funny anecdote, but there’s nothing funny about what’s happening now. Disinformation and conspiracy theories run rampant; attempts to call out fake news are decried by a major party’s presidential candidate as censorship. Knowing who to believe is trickier than ever. And more confusion is on the horizon. What will happen when fake news is accompanied by deepfaked video or counterfeit audio? What will “reputation” mean then? I have no doubt that in seventeen years, this reflection will seem quaint.
The websites I created are long gone, and thankfully so. What I wrote wasn’t meant to endure. I remember resenting that I had to write novels’ worth of text for clients instead of the novels I wished to be writing. But writing nonsense was writing, too. It was practice. Despite everything, the job oriented me toward the writing that I do believe is worth my time: writing that is crafted from contemplation and introspection, that hopes to connect deeply with its reader, that is meaningful to me, as a writer. Lately, this kind of writing feels doomed, too. ChatGPT threatens to take my job—not just my old one, but the one I care about.
Times change—technology arrives—and we change along with it. We can only comprehend those changes with perspective. From here, in the present, I can only wonder what’s to come—how our lives will be transformed. For now, I can do my best to write something true.
You Might Also Like
