How Conservative Activists From Project Veritas Catfished Twitter
By Kashmir Hill
Mo Norai has worked in Silicon Valley for a decade. He’s done stints at Google, Twitter, Facebook, and Apple, but only as a contract worker, meaning he has missed out on the tech giants’ storied perks, benefits and job security. So when he was approached last April by a recruiter from a company called Tech Jobs Box about a full-time job, he was intrigued.
A woman named Kelly Dale contacted him via LinkedIn promising that “salary and benefits would be competitive.”
“It really hooked me,” Norai told me last month.
After a brief phone conversation, Dale said he seemed like a great candidate and set up in-person interviews with her colleague and an investor in the company. Those went well and for four months last year, Norai thought he had a new job. He was in regular communication with his new colleagues, meeting up with them for dinner, drinks, and a baseball game, but they kept pushing his start date back, saying they were securing office space and finalising funding.
But in fact, there was no job. Tech Jobs Box wasn’t a real company. Kelly Dale and the rest of his new “colleagues” were actually operatives for Project Veritas, a conservative investigative group founded by James O’Keefe that specialises in secretly recording people. It’s perhaps best known for catalyzing the downfall of ACORN, a low-income advocacy group that lost its federal funding after Project Veritas released undercover videos of the group’s employees counseling a sex worker and her “pimp” (a disguised O’Keefe).
Project Veritas traditionally targets politicians, government agencies, and media organisations, but decided to go after Silicon Valley last year because of its perceived biases against conservatives. “Big tech companies like Google, Facebook and Twitter have become media monopolies and they are censoring people,” said O’Keefe by phone.
In January, Project Veritas released three videos about Twitter’s content-moderation practices that feature hidden camera footage of nine current and former Twitter employees – one woman and eight men, including Norai. They even secretly filmed Twitter CEO Jack Dorsey by having an operative pose as a homeless person and confront him at a Blue Bottle coffee shop.
The videos don’t contain blockbuster information. The employees reveal that there aren’t a lot of conservatives at Twitter; that Twitter tries to make spammy content less visible on the platform; that many of the sock puppet Twitter accounts banned in the last year posed as Trump supporters; and that Twitter would cooperate, as required by law, in any investigation of President Trump by handing over his private Twitter messages. The most surprising part was a former engineer’s claim that Twitter historically “shadow banned” users. (A “shadow ban” means that a user’s content on a platform can’t be seen and the user doesn’t realise it.)
Project Veritas points out that Senator Ted Cruz cited their videos while questioning tech companies during a hearing about content moderation, terrorism, and Russia in January.
“The individuals depicted in these videos were speaking in a personal capacity and do not represent or speak for Twitter,” said a Twitter spokesperson by email, pointing me to a page that explains how and why Twitter accounts are censored or made less visible. “Twitter does not shadowban accounts. We do take actions to downrank accounts that are abusive, and mark them accordingly so people can still click through and see this information if they so choose.”
While Project Veritas’s findings weren’t particularly shocking, how they were obtained was. Project Veritas didn’t just fake-recruit its targets, it fake-seduced them. Many of the male employees were secretly recorded while on dates at dimly-lit restaurants, sipping wine. Based on the number of times he appears in the videos in different locations and dress, one security engineer, Clay Haynes, appears to have been enamoured enough with the operative pumping him for information to go out with her at least three times. All of the Veritas operatives’ faces are blurred, but you can see his date’s jangly bracelets and long blond hair. It’s unclear just how far the seduction of Haynes went, but they became serious enough to go on a double date to Morton’s Steakhouse with her friend, a disguised James O’Keefe.
“NO ONE should have to experience this,” said Haynes via Facebook message. Haynes, who is still employed by Twitter, ultimately opted not to talk to me at the company’s request.
Beyond the questionable journalistic ethics of exploiting people’s desires for work and love, Project Veritas’s tactics broke the law, says John Nockleby, a professor who specialises in privacy at Loyola Law School-Los Angeles. While consent laws for recording conversations vary from state to state, California is a two-party consent state, meaning you have to tell someone if you’re recording them, or face up to a year of jail time and a $US2,500 ($3,240) fine. “You’re allowed to do video in a public place without getting consent, but not take audio, unless it’s someone like a politician giving a speech to a crowd,” Nockleby told me by phone. “In California, even in a public place, if you’re audio recording without consent, that’s not legal.” In California, even in a public place, if you’re audio recording without consent, that’s not legal.
O’Keefe, who paid a $US100,000 ($129,600) settlement in 2012 to an ACORN employee who sued him over California’s law against surreptitious recording, expressed the belief that his operatives are allowed to record people in public places, like bars, restaurants or a conference room where a door is open.
“We have a number of lawyers who handle compliance for us. California is a two-party state but we can operate in areas where there are no expectations of privacy,” said O’Keefe by phone. “With the Twitter story, we did not break the law. Period.”
In a follow-up email, a Project Veritas spokesperson pointed to an exception in the law for circumstances in “which the parties to the communication may reasonably expect that the communication may be overheard or recorded.”(Norai says the door to the conference room where his interviews took place was closed.)
The story Silicon Valley likes to tell about itself is that it conquered the world by making it more open and connected, and by getting strangers to trust each other. Project Veritas exploited that ecosystem of connection and trust to wage its year-long investigation, turning the tools that Silicon Valley created against it. In a phone interview, O’Keefe declined to reveal how many undercover journalists were involved or how much it spent on the operation, saying only that it “was very expensive” because travel and lodging in San Francisco “was outrageous.” (Project Veritas doesn’t seem to be having money problems; its budget has nearly doubled every year, according to financial filings. O’Keefe says it raised more than $US7 ($9) million in 2017.) He said he couldn’t talk about his group’s methods because the investigation of tech companies is ongoing. Google and Facebook employees should beware. The investigation of tech companies is ongoing. “We still have active investigators out in the field,” said O’Keefe.
“We still have active investigators out in the field, which is why I can’t reveal the techniques used to gather information,” he told me. “I wish these companies would be honest about what they are doing. Unfortunately, we live in a society where there is so much dishonesty that it requires undercover work.”
In other words, deceit requires more deceit. I was able to unearth some of the group’s methods of deception: It sent targets messages on LinkedIn and dating apps, built false identities on social networks, and operated a fake start-up out of a WeWork blocks from Twitter’s headquarters.
The larger revelation of the project is not that Twitter is biased against conservatives; it’s that Silicon Valley has given strangers who bear you ill will all the tools they need to infiltrate your life. Our identities are scattered across the web on multiple platforms, giving people countless ways to make contact with us as well as dossiers of what we do, whom we like, and where we go.
The Twitter targets, who had been secretly recorded at parties, in bars, on first dates, and during job interviews, only found out about the deception on January 11, when the first Project Veritas video, titled “UNDERCOVER VIDEO: Twitter Engineers To ‘Ban a Way of Talking’ Through ‘Shadow Banning,’ Algorithms to Censor Opposing Political Opinions” went live.
“You’re on the national news talking about Twitter,” a friend told Mo Norai via Facebook message after the first video was published. Norai had no idea what she was talking about at first. “You know I don’t work at Twitter anymore, right?” he sent back.
Then he went to Project Veritas’s site and watched the video. “Oh fuck, this is bad,” he thought.
In the videos, Norai is dressed formally, wearing a tie and a black suit vest, his long hair pulled back in a ponytail. Norai’s experience at tech companies is in content moderation — deciding, essentially, what speech needs to be taken down. He’s removed spam from iTunes and reviewed posts as a contractor for Facebook. He was hired as a contract worker at Twitter after the Paris attacks in November 2015 because terrorism results in increased offensive postings on social networks. “We were warned it was going to be grisly,” Norai told me.
He was there for just three months, and was gone by March 2016, months before Trump became the Republican presidential nominee. Having “Content Review Agent” for Twitter on his LinkedIn résumé, though, made him a target of Project Veritas. Vaporware,a successful strategy for many a start-up in Silicon Valley, also works well for deceptive investigative operations.
Norai had two interviews with the fake start-up, which was allegedly going to make content recommendations to people based on their location and buying patterns. He says he typed the company’s name into a search engine and didn’t find anything. (Vaporware, a successful strategy for many a start-up in Silicon Valley, also works well for deceptive investigative operations.) “It seemed weird but I met the people,” Norai told me. “They said they wanted to start a start-up. It made sense it wasn’t in existence yet.”
Both his interviews were in May. Kelly Dale, the supposed recruiter, said the office was in downtown San Francisco and gave Norai an address on Mission Street. When Norai got there, it turned out to be a WeWork co-working space, which, again, is not that strange. Lots of early-stage start-ups are based out of WeWorks around the country, because they offer easy plug-and-play office space. The WeWork-Civic Center that Project Veritas used rents office space for as little as $US780 ($1,011) per month.
I went to the WeWork last month – it’s a seven-minute walk from Twitter’s headquarters, in a stylish building on an otherwise seedy street in San Francisco’s Tenderloin district – and asked for Tech Jobs Box. The receptionist recognised the name, but said they weren’t there anymore. She wouldn’t say how long they were there or how many people came in to meet with them. WeWork declined to answer any questions about the fake company that rented space from it.
Norai’s interviews were with a doughy-faced guy who introduced himself as Eric Williams. A grey-haired “investor” named Dan, who Norai later discovered was Dan Sandini, a known operative of Veritas, came to the second interview, the one posted online. Dan and Eric said they lived on the East Coast but wanted to recruit talent in the Bay Area.
Norai thought the interview questions were odd. They asked him theoretical questions about how he might have dealt with tweets from aggressive Trump supporters.
“The political questions were strange but I just thought it was because they weren’t from here,” said Norai, a life-long Californian who grew up in the East Bay.
Norai hadn’t worked at Twitter during the bulk of the 2016 campaign, nor had he dealt with Trump-related content moderation. But his theoretical comments were presented as if from experience. Project Veritas’s habit of presenting comments out of context is part of why people question the credibility of their videos. James O’Keefe says Project Veritas recently formulated a code of ethics that includes “not breaking the law” and not deceptively editing videos or “putting words in people’s mouths.”
After Norai’s interview ended, Eric and Dan took him out for celebratory cocktails at two swanky San Francisco bars, Local Edition and the View. They said Norai was hired. At the View, Norai took out his phone to Snapchat the moment. Dan protested and asked Norai not to put it online. “I just thought he didn’t want to look silly because I gave him cat ears,” said Norai. ”Current or prior Facebook or Twitter experience preferred.”
When Eric came back to town in June, he invited Norai to a Giants game with two other men, who said they were Department of Justice lawyers. The four men sat behind home plate, and Eric asked for Norai’s help in recruiting other people for the start-up. Afterwards he sent him job descriptions for two positions: “software engineer, machine learning” and “spam operations.” “Current or prior Facebook or Twitter experience preferred,” the listings read, suggesting Facebook may get the Veritas treatment next. Norai put Eric in touch with two friends who had experience at Twitter and Facebook, respectively.
The night ended at bars again. After drinks at the Starlight Room, Norai got on the train home to the East Bay. “It felt like we’d established a friendship,” said Norai.
He never saw any of them again. At some point, Norai tried to email Kelly Dale and it bounced. Eric told him she’d gotten married and had a new email address, which made little sense given that her email was “email@example.com.” A few months later, Eric’s number stopped working. Dan told him Eric had been let go. Still jobless months later, Norai began to wonder if it had been some kind of scam. He decided to take a contract work position at Facebook last September.
When the Project Veritas video came out in January, he found out exactly what kind of scam it was. The videos didn’t include the two friends he put Eric in touch with.
“I hope it’s over. I really hope it’s over,” Norai told me.
In January, an investigator from Twitter reached out to him to ask if he had known he’d been filmed and asked whether he’d been contacted on Tinder or other dating apps, as others had been. While Norai says he mostly worries about potential professional damage, he also bears psychological scars from the episode. He says he’s paranoid now about meeting new people, and worries when a phone is out during an interview because it could be secretly recording him. He didn’t feel comfortable with me until I showed him my driver’s licence.
He asked me if I’d ever done anything like this before, which was an uncomfortable question, because I have. A couple years ago, I created a nonexistent business — the Freakin’ Awesome Karaoke Express, or F.A.K.E. — and made it a website as well as accounts on Twitter, Facebook and Yelp, in order to explore the fake reputation economy. I do believe that some stories can only be told well by going “undercover,” and while I hate to admit it, part of me is impressed by all the work that Project Veritas put into its Twitter exposé. Many a journalist knows it can be hard to get people who work at tech companies to talk. But their specific tactics were deplorable and cruel. They didn’t go undercover to find out how an industry actually works; they went undercover to exploit people’s vulnerabilities in the pursuit of information that could have been unearthed with a more traditional journalistic approach.
O’Keefe disagrees with that sentiment. “It’s hard to get [tech companies] to be honest on the record. You call them as a journalist and they say they don’t do it. We got tipped off to shadow banning and set out to confirm it. And we basically confirmed it,” he told me by phone from an Uber. “Sometimes that’s when you’re most honest, when you’re being recorded in private.”
Twitter disputes the honesty of the investigation. “We deplore the deceptive and underhanded tactics by which this footage was obtained and selectively edited to fit a pre-determined narrative,” said a spokesperson by email. “Twitter is committed to enforcing our rules without bias and empowering every voice on our platform, in accordance with the Twitter Rules.”
Since the Twitter investigation was published, Project Veritas hasn’t done much to cover its tracks. Kelly Dale’s LinkedIn profile is still up, with over 200 followers and claims that Dale worked at CUNY and went to Pace. Neither have records of her for obvious reasons: She doesn’t exist. According to Nexis’s people search engine, there are no Kelly Dales who live in Massachusetts. The profile photo looks to me like it belongs to Allison Maass, who did not respond to a media inquiry sent via LinkedIn. Maass left Project Veritas last year to work for Circa, an online media group owned by the media behemoth Sinclair Broadcasting, which famously requires its television stations to air conservative news daily, and is currently trying to expand its footprint by acquiring Tribune Media. Project Veritas and Sinclair Broadcasting have collaborated in the past.
Tech Jobs Box has an anemic LinkedIn page with banal stock art and a link to a website that doesn’t exist. It only recently established a Twitter account, joining the network in October.
Leaving behind the traces of abandoned identities is pretty common for Project Veritas, said Lauren Windsor, an executive producer at the web show Undercurrent who was working at Democracy Partners when the group of political consultants was infiltrated by Project Veritas in 2016. After Democracy Partners realised it had been duped and secretly filmed by an intern named “Angela Brandt,” who was in fact Allison Maass, Windsor began preserving evidence for a million-dollar lawsuit against the group. She realised it would be useful to publish that evidence – photos and details of faked identities — to help others avoid the same fate. In January, she published an online dossier with the photos and known aliases of over 150 Project Veritas operatives.
When shown a photo of “Eric Williams,” Windsor recognised him from another Project Veritas operation and sent me photos of him filming activist group Voces de la Frontera in Milwaukee in 2016.
Many of Project Veritas’s recent stings haven’t gone very well. Investigations into the Washington Post and the Soros Foundation were spectacular failures. But even if Project Veritas’s methods are crude and “amateurish” as the New Yorker‘s Jane Mayer put it, Windsor warns that it is a force to be reckoned with.
“They have gotten a lot of funding. They have been on a recruitment surge to try to staff up before 2018,” she said by phone. “I think they’re a real threat to media and political organisations, and, most importantly, to candidates.”
O’Keefe echoes that. “We have an untold amount of people inside organisations and they don’t know that,” he told me. “We are going from the smash-and-grab to the long con.” We are going from the smash-and-grab to the long con.
Windsor says the group’s techniques are “inhumane.” “They see themselves as being at war with the Left but the [people they secretly film] are real people despite how much Project Veritas dislikes them politically,” she said. “A lot of people [who are infiltrated and secretly filmed] have described the experience to me like psychological rape. It felt like that to me. They’re destroying lives and it’s really important to expose their political espionage as much as possible.”
Windsor has been training liberal groups how to spot a Veritas operative, mainly by thoroughly researching the online identity of anyone who offers free help or dangles the offer of donor money. If their online footprint is thin, it can be a clue. But there are, of course, legitimate reasons why someone might have a thin online footprint, so it’s not a sureproof solution.
I don’t actually have any sureproof solutions for you. We could ask social media companies to try to protect us from deceivers by authenticating people’s identities – Tinder could require users’ driver licenses; LinkedIn could ask for business records in order to list a new workplace — but that would simply invite a different sort of dystopia. We’re never going to be able to completely eradicate fake bots on Twitter or faked videos or con men on dating sites. Trolls will always find a way.
Maybe just keep in mind that, despite Silicon Valley marketing to the contrary, being in a more open and connected world has downsides. Project Veritas may not be particularly great at spycraft, but with enough manpower and enough online information at its disposal and enough hidden cameras in enough places, it will compromise people and organisations in damaging ways, fairly or unfairly. It’s the real-life manifestation of the faceless algorithms and unknown data brokers that scrape our profiles for information and constantly sit in the backgrounds of our browsers tracking everything that we do. Instead of using it to serve you Instagram ads for shoes, they’re trying to take down your organisation. You will just be a casualty along the way.