Joan Donovan, one of the world’s leading experts in misinformation, was dying to set the record straight. On a brisk November night, she told me a story about why she’d left Harvard University. It was captured, she claimed, by a corporation she had loudly criticized, one with far too much power over our democracy: Meta.
Donovan had been preparing for months to air this accusation in public, and I’d flown to Boston to interview her before the big day. At the moment, she had just heard through her lawyer that Harvard wanted to talk. “What are they going to offer me, $5 million?” she mused as we sat in a cafe. She wore a leather jacket over head-to-toe black, and a whistle dangled from her neck. “How am I going to feel about that money if I don’t tell the truth?”
We’re sorry, something went wrong.
We are unable to fully display the content of this page.
This is most likely due to a content blocker on your computer or network.
Please allow access to our site and then refresh this page.
You may then be asked to log in, create an account (if you don't already have one),
or subscribe.
If you continue to experience issues, please contact us at 202-466-1032 or help@chronicle.com.
Joan Donovan, one of the world’s leading experts in misinformation, was dying to set the record straight. On a brisk November night, she told me a story about why she’d left Harvard University. It was captured, she claimed, by a corporation she had loudly criticized, one with far too much power over our democracy: Meta.
Donovan had been preparing for months to air this accusation in public, and I’d flown to Boston to interview her before the big day. At the moment, she had just heard through her lawyer that Harvard wanted to talk. “What are they going to offer me, $5 million?” she mused as we sat in a cafe. She wore a leather jacket over head-to-toe black, and a whistle dangled from her neck. “How am I going to feel about that money if I don’t tell the truth?”
In buttoned-down academe, Donovan is an outsider: a two-time college dropout from working-class Massachusetts, a lesbian punk anarchist who embedded in the Occupy movement and emerged with a sociology dissertation. So prescient was her understanding of today’s fractured media landscape — and how lies and hate speech thrive inside it — that Harvard’s John F. Kennedy School of Government recruited her in 2018. Throughout the Trump presidency, a pandemic, an insurrection, and a societal panic over disinformation, she gave countless media interviews, testified before Congress, and emerged as a trusted guide for frightening times. A defender of a republic and a press under siege. A truth-teller in a sea of liars.
Then, she said, Harvard kowtowed to top brass at Meta — the trillion-dollar social-media giant that owns Facebook, Instagram, and WhatsApp. Donovan was not shocked: she had seen the academy corrupted before, by Big Pharma and Big Oil. Now her university, influenced by its financial and personal ties to Meta, had eliminated her role and the team she led, alleged a 248-page document she released in December and filed with Harvard, the U.S. Department of Education, and the Massachusetts attorney general. Whistleblower Aid, a legal nonprofit representing her, demanded investigations. (A Meta spokesperson declined to comment.) Even though Donovan by then had a new job, at Boston University, she told me, “I would be complicit if I kept my mouth shut.”
Her document was not a lawsuit, but a first-person declaration of how the world’s wealthiest university, hamstrung by its “significant conflict of interest,” allegedly mistreated her up to her departure last summer. It had, she claimed, taken the copyright to her book. Stolen her plans to publish confidential Facebook documents. Blocked an event she was required to host. And jeopardized the livelihoods of her staff, the now-disbanded Technology and Social Change Project, which was part of the Kennedy School’s Shorenstein Center on Media, Politics, and Public Policy.
I would be complicit if I kept my mouth shut.
Joan Donovan
The Washington Post, The Boston Globe,NBC, NPR, and dozens of other outlets (including this one) carried her claims, along with Harvard’s insistence that the team ended for a purely administrative, only-in-academe reason: She was a staff member, not a faculty member, so as a matter of policy, she could not lead research indefinitely. Laurence H. Tribe, the Harvard constitutional-law scholar, declared that the university “has to be investigated from top to bottom.” This, Donovan told CNN, was “a knock-down, drag-out fight for my academic freedom.” Her eyes welled with tears.
Here was a narrative with the kernels of some undeniable truths. Meta doesfunnel money into higher ed; Harvard is cozy with the 1 percent. But a believable story is not necessarily a true one. Donovan presented no firsthand evidence that Meta was behind her ouster. And when I tried to get to the bottom of what actually happened at Harvard, a different narrative emerged from interviews, documents, recordings, texts, and emails.
Eleven Technology and Social Change Project ex-members and Shorenstein staffers told me they had seen no evidence that Meta exerted pressure on Donovan’s team or that its influence is what ended it. Several of Donovan’s other claims about her time there are misleading, untrue, or contradicted by people directly involved. Some former colleagues say they no longer trust the scholar they once admired.
On my last evening in Boston, I asked the media-manipulation scholar how confident she felt in the narrative she was about to release into the world. Could this all just be a result of garden-variety academic bureaucracy?
“Well, if it is,” Donovan replied, “let’s find out.”
“Fuck you! Pay me!” Donovan screamed over wailing guitars. “I don’t work for free!” We were driving through her hometown of Saugus, a small suburb north of Boston, and blasting her band, RANT. (Full name: Rebel Against Nazi Tyranny.) The scholar harbors a deep sense of rage at the injustices of the world; hardcore punk is her outlet. “Screened in by opinions, screened out by algorithms,” she shrieked over the speakers of her electric-blue Dodge Challenger. “All of our institutions come apart.”
Donovan still lives near where she grew up in Massachusetts. She delivers witty one-liners with flashes of a Boston accent, a wicked cool surfacing every now and then. Her house, which she shares with two cats, is lined with concert posters and photos of herself moshing with a mic. A constellation of tattoos mark her allegiance to the straight-edge scene — no drugs, no drinking — and groups like Alkaline Trio and Jawbreaker. When she was in her 20s, her band was kicked off Harvard’s campus for sets that got a little too rowdy. “There’s quite a few shows,” she recalled, “where people were very nude.”
That, at the time, was the closest she thought she’d get to Harvard. Neither her mother, a phlebotomist, nor her father, a postman, have college degrees, and while she did well in high school, she was better known as the class clown who could solve her friends’ personal problems. An aunt signed off on her student loans, and Donovan hopped from Northeastern University to the University of Massachusetts at Boston, with time off in between to deal with her parents’ divorce. But she was bored with the business classes she was taking, and dropped out, feeling lost.
In 2004, when George W. Bush was re-elected, Donovan’s drummer at the time persuaded her to move to Montreal. It was cool and cheap, and, for a young queer woman, it was “nice to live somewhere where existing wasn’t so contentious,” she told me. (Donovan is not registered with any party and does not vote, which she says is to protest the United States’ pro-war stance. She made an exception for Joe Biden in 2020.) At Concordia University, she finally wrapped up her undergraduate degree, nearly a decade in the making, in sociology, a field that unexpectedly captured her imagination. It gave her a framework for all that righteous fury. “I loved the idea of thinking about the world and the mechanics of society and understanding politics for what it is, which is a struggle over power and resources,” she told me.
Donovan’s professors encouraged her to attend graduate school, a pivot that felt surreal to someone who’d previously thought that only medical providers were “doctors.” For her master’s, also at Concordia, she studied how mental illness was stigmatized, a subject she’d observed firsthand as a counselor for mentally ill adults in Boston. For her Ph.D. in sociology and science studies at the University of California at San Diego, she thought she’d keep up that research, and perhaps end up as a mental-health advocate.
ADVERTISEMENT
But in 2011, a crowd of protesters camped out at Los Angeles City Hall, angry at the Great Recession and the enormous wealth disparities it exposed. “I decided I was going to go and see what it was all about,” said Donovan, who was living nearby at the time. “I didn’t really go home after that.”
For a budding sociologist, Occupy was a fascinating object of study: a movement that mobilized thousands without anointing individual leaders. Those activists were stitching together ways of talking on the fly, using existing tech — conference calls, Google, Twitter, Facebook — as well as open-source tools built from scratch. Donovan threw herself into the dual role of an activist-academic, alternating between organizing calls and writing her dissertation about Occupy’s communication infrastructure. “Occupy really changed activism, especially in the U.S.,” she told me. “It shook up all the networks of activists, in the way that labor organizers were talking with anarchists were talking with anti-eviction activists were talking with military vets.” Martha Lampland, her dissertation chair, remembers some colleagues saying that Donovan shouldn’t be studying something she was involved in. “I didn’t take them seriously,” Lampland told me. Even then, she knew her student “was always going to be in the center of all kinds of things.”
Donovan also led protests of her own, against tuition hikes. (“You would think with all this technology, school would be absolutely free by now,” she told me.) When a student publication drew heat for making racist jokes, she rallied friends to make fake issues “apologizing.” Kelly Nielsen, a friend from her program, credits Donovan’s punk ethos for her gift at “spotting bullshit and being able to confront it in really creative and funny and often pretty damning ways.” She is, he added, “totally fearless.” “I always sort of fell more on the side of just being interested in how things work,” said another classmate, David Pinzur, “and Joan was always the one who was much more concerned with ‘What is the world like?’ so that we can then go out and make it better.”
As a postdoctoral fellow at UCLA, Donovan was again early in understanding how technology was shaking up the culture. Acting on a tip passed on by her supervisor, she combed through Stormfront, the infamous white-supremacist forum, where posters were taking new consumer DNA tests to prove their “whiteness.” Some, to their dismay, were getting back strains of African or Ashkenazi Jewish heritage. (They coped in part by calling the tests bogus.) Aaron Panofsky, the sociologist who led the study with Donovan, marvels that she had the intuition to corral all the researchers into the same room, rather than work separately, so they could band against the vitriol they were reading. “It was a transformative experience for me to watch Joan make this realization, make this pivot, and organize this group of students into this kind of research collective,” Panofsky told me. White supremacy, at the outset of the project, was not mainstream news. Then the pair presented their findings at a conference in August 2017, days after members of the growing alt-right gathered offline for a violent rally in Charlottesville, Va.
There are a lot of people who are doing really good research; there are very few people capable of fighting the fight.
Craig Newmark
The pledges of the techno-optimist era — move fast and break things, connect the entire world — were suddenly looking like threats made good. After Donald Trump won the White House in 2016, BuzzFeed News reported that on Facebook, made-up election-news stories had more engagement than real ones. Mark Zuckerberg, the company’s chief executive, initially said it was “pretty crazy” to think that fake news had influenced the vote, then backtracked. Around the world, Facebook faced accusations of amplifying pro-government propaganda and fueling genocide. Experts who understood the digital hellscape were in desperate demand.
Donovan had by then landed a research gig at Data & Society, a New York think tank that studies technology’s social implications. “I wasn’t hot off my postdoc getting tenure-track appointments,” she told me. So when she showed up at a conference at Harvard, the last thing she expected it to lead to was a job. “I just remember this feeling of being completely overwhelmed,” she said. She joined the Kennedy School’s Shorenstein Center at the end of 2018, and was promoted to research director in 2020. “She’s both a good researcher and smart about talking about it,” said Craig Newmark, the Craigslist founder and one of the philanthropists who funded Donovan’s work at Harvard. “There are a lot of people who are doing really good research; there are very few people capable of fighting the fight.”
Most scholars’ top priority is getting these kinds of findings into journals. “I love the battle of peer review,” Donovan told me. But “I make knowledge for the people. And that doesn’t mean that I don’t follow methods and then I’m quick to assert things that the data doesn’t show. It just means that I have to be more imaginative with the projects I take on and think, ‘Is this going to be relevant two years from now, when we actually go to publish?’” For years, she’s talked to the public through op-eds, podcasts, lectures, and the platform formerly known as Twitter, where she has more than 45,000 followers.
Being extremely online makes her a go-to source for tech journalists. In 2018 she noticed that prominent white nationalists on YouTube were making thousands of dollars from a feature that let commenters — in their case, racist ones — pay to ask questions. When YouTube was told and declined to act, she recalls, she turned to Craig Silverman, BuzzFeed News’s then-media editor (and my former colleague). After he and a coworker wrote about the phenomenon and The Wall Street Journal followed suit, a far-right channel came down. Donovan was also early and relentless in sounding the alarm about the coronavirus “infodemic” in World Health Organization briefings, scientificjournals, and U.S. House hearings. Members of Congress urged Biden to add her to his Covid task force. “I felt unbelievably useful for almost the first time in my life,” she told me.
Every night for years, Donovan binged on white-supremacist podcasts and videos, listening for new catchphrases and ban-evasion tactics. (How, I asked, does she avoid getting red-pilled? “Well, if there’s one thing you’ve probably deduced about me,” she deadpanned, “is I have some strong opinions.”) With her ear to the ground on the eve of 2021, she warned her team: “You should all be aware that January 6 is going to be a big, big deal.”
Heavily armed Trump supporters proceeded to storm the Capitol, determined to overturn an election they baselessly believed was rigged. Donovan watched their live streams and recognized far-right leaders, like Alex Jones and Nicholas Fuentes, on the spot. She translated every shouted meme. “She should have been on CNN, being like, ‘This is what this person is talking about,’” Emily Dreyfuss, then a senior managing editor on the Technology and Social Change Project, told me. That night, she suggested to Donovan that she should write a book. Published the following year, Meme Wars: The Untold Story of the Online Battles Upending Democracy in America ended up being a joint effort between Donovan, Dreyfuss, and their colleague Brian Friedberg. It recounts a decade in which social media transformed real-life political activism, for better and worse, starting with Occupy and ending with the “red-pilled right” converging to #StoptheSteal.
In the crowded field of internet studies, most of the oxygen gets sucked up by research that draws on enormous data sets. Donovan prefers the sociologist’s methods of direct observation: interviewing, surveying, embedding in communities, chronicling how people interact with technology and one another. “You can drain the lake or the ocean, and there’s plenty of fish,” as she has put it. “Or you can build a better fishing rod.”
The perils of draining the ocean became clear in 2018, when a political consulting firm was revealed to have harnessed the data of up to 87 million Facebook users — improperly collected by an academic — to help Trump target voters. The social-media conglomerate has since invested heavily in research about how its products affect democracy, to the discomfort of more than a few researchers. Facebook founded Social Science One, a Harvard-based organization, only to delay sharing data with the partnering scholars for more than a year and then admit that it had major errors.
During the 2020 U.S. election, a team of Meta researchers partnered with outside scientists to study Facebook and Instagram’s impacts on tens of thousands of users’ political attitudes and behaviors. The company didn’t pay the academics or have final say over their findings. But the academics weren’t allowed to handle the users’ raw activity data. And when Meta described the first four studies, published last summer, as refuting concerns that its algorithms polarize people, the academic authors objected. Michael W. Wagner, a journalism professor at the University of Wisconsin at Madison who independently observed the process, told me, “They’ve done really careful and rigorous and replicable social-scientific research, but they haven’t done so in a way that completely makes them independent from the social-media platforms they’re trying to understand.” Yet Joshua A. Tucker, a political scientist at New York University who was a leader of the project, said that researchers have to choose between working with the companies, working without them, or lobbying for data-access mandates. Each approach has flaws, he said, but forsaking all of them is not an option: “This is so important a topic that we can’t do that.”
Facebook tried courting Donovan in the past. Amber Heffernan, an employee working on “global strategic initiatives,” contacted her multiple times at Harvard, according to Donovan, at one point asking what she would study with a blank check (she’d go live on an island, she replied) and how Facebook data could enhance her work (it wouldn’t). (Heffernan did not return a request for comment.) No data, no funding, no nondisclosure agreements: These are Donovan’s lines in the sand when it comes to Big Tech. “I am not interested in improving the speed and efficiency of tech companies’ workflows,” she told me.
Last fall at a hotel in Montreal, Donovan, wearing her trademark leather jacket, told a gathering of internet researchers that their overreliance on companies’ generosity was backfiring. Meta was on its way to killing a tool that tracked the reach of Facebook posts, and Elon Musk’s Twitter was charging scholars significantly more for data access.
ADVERTISEMENT
Onstage, Renée DiResta, technical research manager at the Stanford Internet Observatory, argued that independent, valuable research could be done with platform-provided data. “I don’t trust,” Donovan retorted, “that data given to a network of researchers hasn’t thoroughly been vetted by Facebook and cherry-picked.”
Rebekah Tromble, an associate professor at George Washington University who’d been part of the team that used Meta’s data to study the 2020 elections, said she considered the partnership worth the potential downsides. “We have learned the ways the platforms work, we’ve learned what it is that researchers should be asking for more of,” she said, and that knowledge could inform how lawmakers regulate data access.
Donovan was unimpressed. Meta “took your reputation, and they laundered it to spread their propaganda, right?” she replied. Afterward, she told me, “I would be doing a disservice if I just kept my mouth shut and agreed that this was fine.”
That candor endears her to some of her fellow scholars. “She’s a breath of fresh air in academia,” Britt S. Paris, an assistant professor of library and information science at Rutgers University, told me. But others have accused Donovan of being a bully. In 2021 she was on the editorial board of the Kennedy School’s misinformation-research journal when it published a controversial study. A slavery-reparations advocacy group, the paper found, had committed “disinformation creep.” The journal later retracted it, saying that reviews of its underlying tweets invalidated the findings. But a Hunter College sociologist — who was not involved with the research — accused Donovan of “driving” the retraction as part of a “nefarious whisper campaign” against the junior Black scholar who led the study. “You have used your considerable, if provisional and untenured, power at Harvard to punch down at someone who is an immigrant, Black, and a single mother,” the professor, Jessie Daniels, wrote in a letter in October 2022. (She attributed the allegations to unnamed “multiple sources.”) Donovan responded with a cease-and-desist letter. She denied to me that she’d staged a “whisper campaign” or single-handedly led the retraction, pointing out that she was one of dozens of editorial board members. Daniels declined to comment on the letter.
While the episode was causing a stir behind the scenes, Donovan’s celebrity was on the rise. At a star-studded disinformation conference in 2022, hosted by The Atlantic and the University of Chicago, she shared top billing with Barack Obama, Sen. Amy Klobuchar, the tech journalist Kara Swisher, and Maria Ressa, the Nobel Peace Prize-winning investigative reporter. “It was just one of those moments where you realize, Holy shit, I made it,” she told me.
Several dozen of the Kennedy School’s major donors were gatheredon Zoom, listening to Donovan talk about her research, when the moderator brought up the bombshell Facebook news. It was October 2021, and The Wall Street Journal had recently reported, based on thousands of documents leaked by an employee, that the company was failing to fix aspects of its platforms that it knew were causing harm. Instagram, for example, was worsening body-image issues for a sizable percentage of teen girls.
As Donovan would later recount in her whistle-blower declaration, she told the group, known as the Dean’s Council, that she had obtained her own copy of the confidential files — “the most important documents in internet history.” She expressed alarm that Facebook, which had just rebranded as Meta, was harming democracy. Among those listening was Elliot Schrage, a graduate of Harvard College, Harvard Law School, and the Kennedy School. He was also, until 2018, Facebook’s oft-maligned head of communications and public policy. The two of them then had a conversation that — Donovan has repeatedly asserted — set in motion her downfall.
When the moderator opened up the floor for questions, according to Donovan’s disclosure, Schrage “monopolized the discussion by accusing me that my reading of the documents was inaccurate and that he disagreed with all prior discussion about Facebook.” Donovan claims that she “tried to answer Mr. Schrage’s allegations, but he kept speaking out angrily.” His conduct was “so overwhelming and disruptive,” she claims, that someone else “was forced to raise her voice in an attempt to try and calm Mr. Schrage.” The mood was “tense, awkward, and embarrassing for everyone involved.”
But a recording of the meeting contradicts that account. In a video the Kennedy School shared with me, Schrage is called on to ask a question, and begins by saying, “I disagree with a tremendous amount of the characterization and analysis that’s been provided, but that’s not the topic here.” He does not bring up the leaked Facebook files. He does ask Donovan how she defines misinformation, and whether television-news networks should be punished for reporting it. He also asks whether a company like Facebook should be obligated to take down a media outlet if the Philippine government considers it to be spreading falsehoods. In all, Schrage speaks for three minutes. Donovan responds uninterrupted for five-and-a-half, mostly to his first two questions.
“Thank you, Joan,” another council member says, then asks Donovan to discuss the subject of social media’s financial incentives. Two other attendees go on to raise questions that she also answers. Schrage does not speak again before the session ends, the recording shows.
This was the exchange that Donovan says jeopardized her career. That night, “I was worried I was going to lose my job based on Mr. Schrage’s anger,” her declaration says, and nearly a year later, when she learned that her team would be ending, she would think back to “when Elliot Schrage became enraged over my statements and possession of the Facebook files.”
Andrew Bakaj, a lawyer for Donovan, suggested that the video I’d seen could be incomplete or edited. (A Harvard Kennedy School spokesperson says the recording was not altered.) Bakaj also noted that at least some other participants also found the conversation to be confrontational. After the meeting, Donovan texted the moderator to ask if she should worry “about the way Schrage got mad at me.” The moderator replied: “I think we should have worried if he DIDN’T get mad.”
There are millions of resources lost to mitigating misinformation-at-scale, where the cost of doing nothing is even worse.
Joan Donovan
But Douglas W. Elmendorf, the dean of the Kennedy School, told me that he would have intervened if he’d thought Schrage was reacting in “an angry or a pressure-driven way.” Schrage calls Donovan “an unreliable narrator.” “I reject Donovan’s characterization of that meeting or my motivations,” he told me in an email. “In our exchange, she simply wouldn’t or couldn’t answer very basic questions about her work and I think this understandably annoyed her.”
In Donovan’s corner of the world, scholars, tech employees, journalists, and policymakers frequently argue over what it is, exactly, they study and how big a problem it is. They question whether mis- and disinformation should be the terms du jour when propaganda, satire, rumors, fake news, and hoaxes date back millennia. They debate whether the obsession with “misinformation” (generally defined as false and misleading information) and “disinformation” (when shared to mislead) obscures the sometimes genuine difficulty of defining “information.” (Did the pandemic start in a lab? Scientists have been fighting for four years over whether that’s a plausible hypothesis or a conspiracy theory.) “Misinformation,” some say, is now just code for views one disagrees with. (Right-wing figures have harassed Donovan online and off, and accused her of perpetuating a “censorship-industrial complex.” Other researchers have faced even greater scrutiny, in the form of congressional subpoenas and public-records lawsuits.)
Quantifying the effect of misinformation is even harder than defining it. In the debate over why people fall for conspiracies, some scholars say that too much attention is paid to social media’s role and not enough to other factors, like government officials who make false claims on prime-time TV. Studies have failed to reliably find a direct causal relationship between viewing online misinformation and changing specific behaviors, such as switching voting positions. But to Donovan, Facebook’s ability to disseminate falsehoods at unprecedented scale has obvious consequences. When vigilantes take up arms in the wake of online rumors about “antifa” invaders, when people read on their feeds that vaccines are microchipped and voting is rigged, other members of the public — law-enforcement officials, doctors, journalists, election workers — spend time debunking and reassuring. “There are millions of resources lost to mitigating misinformation-at-scale, where the cost of doing nothing is even worse,” Donovan has written. She is among those advocating for “a public-interest internet,” one where social-media feeds would be required to contain “timely, relevant, and local” news curated by librarians.
ADVERTISEMENT
Ten days after the Zoom call, Elmendorf asked Donovan by email if they could meet to discuss issues like these, spurred by “questions raised by the Dean’s Council and my own limited reading of current events.” He wanted to know how her research justified her criticism of Facebook. And how did she define misinformation in situations with “no independent arbiter of truth”?
That phrase caught Donovan’s attention. Meta executives often say their company cannot be “the arbiter of truth.” “It was obvious to me,” her disclosure says, “that Facebook, either directly or through intermediaries, was communicating with him about my research.” Elmendorf, for his part, says the term “doesn’t strike me as a trademarked phrase of any sort.” He told me he doesn’t remember if Schrage contacted him after the meeting, and the former Facebook executive wouldn’t say whether he did when I asked. In general, Elmendorf says that he talks “periodically” with Schrage, who has belonged to the Dean’s Council for years, and that their relationship is “personal” but “not close.”
Writing back to the dean, Donovan explained that she wasn’t in the business of passing judgment. “We look for behaviors where individuals or groups have created fake accounts, or bought advertising, or have sought to wedge public issues in some underhanded way by hiding their identity, affiliations, or intentions,” she wrote. “In this way, we do not generally speak about what is good or bad for a society, but rather what is true or false about a specific public event.”
The following January, she elaborated on this approach in the meeting Elmendorf had called. “That was not a judgment meeting, it was an information meeting for me,” he told me, the kind he says he sometimes has with staff to learn more about what they’re working on. This one left him thinking that her research was “very interesting.” But Donovan thought that he’d tried to “poke holes in the work,” she writes in her declaration. It was an abrupt-seeming turn from a little over two years prior, when he’d sent her a message praising “all the wonderful work you’ve been doing.”
That work was being done by a staff member and adjunct lecturer, not a faculty member — a distinction that Elmendorf says was mattering more and more to him. Harvard policy states that research must be led by faculty, and a Shorenstein Center director had acted as Donovan’s faculty leader until he left in 2019, Elmendorf said. Starting in 2020, with the dean’s permission, a computer-science professor had agreed to share leadership of Donovan’s projects. But neither she, nor the center’s current director, nor other faculty with relevant expertise wanted to be responsible for all of Donovan’s work in the long term, Elmendorf told me. The Technology and Social Change Project at its peak also had a relatively large staff of about 30 full-time, part-time, and temporary employees, which made it that much more of a time demand.
“It definitely seemed worth it to me and to others to try to give this a chance to be taken up by a different faculty member,” Elmendorf told me. But he added that “the dean is not given the flexibility, for good reason, to just ignore the policy.” When the professors John Ruggie and Ash Carter died in 2021 and 2022, respectively, and Wendy Sherman left in 2021, their programs also ended because no other faculty members took over. “None of these cases is easy,” the dean told me, “but it is a regular, recurring feature of what we do.”
Donovan didn’t seem to grasp the precarity of her employment. (“I was just as competent as any other person at Harvard, and I deserved the same amount of dignity as they gave everybody else,” she told me. “So this notion that somehow I was ‘just staff’ makes sense to them, but to me, they’re also ‘just staff.’” She added, “I’m not particularly enamored by titles.”) On August 24, 2022, Elmendorf told her that the Technology and Social Change Project would be wound down, and until then, she could not start new projects, grow the team, or raise funds. He later specified that she could have until June 2024, which he told me was probably chosen to coincide with the end of the academic year. Donovan’s contract was originally set to end in December 2024.
Citing the policy that only faculty can run research, Elmendorf told Donovan, “I want to remind you that you’re staff here,” according to her, and said she had become “too prominent.” (A Harvard spokesperson says that it “did not try to curtail Joan’s public pronouncements and she was frequently quoted in the media until her departure.”) Finally, she alleges, Elmendorf told her that she could apply for a tenure-track job he was “going to” create. The dean told me that he’d said he was “hoping” to create it, though the school ended up not doing so. “We worked hard to give her and others who worked on these projects time to make other plans,” he said.
Stunned, Donovan first tried to relocate her team elsewhere at Harvard. But talks with two centers fell through. (She accuses Elmendorf of “sabotaging” those efforts, which he denies. “From the Kennedy School’s perspective, Joan’s moving to some other place at Harvard would have been the smoothest, most straightforward way to wind down,” he said.) In conversations during the fall of 2022, according to two former team members, Donovan blamed Elmendorf and Harvard — not Facebook. (Donovan told me that at this time she was starting to learn about the relationships between people at Harvard and Meta, and wasn’t sharing that information widely.) She also floated the theory that she was the target of bias. Texting a teammate in September 2022, she confided that she’d been advised to file a complaint with Harvard over sexism and homophobia. (She told me that she never filed one but that she “might have been thinking it through.”)
Within the Shorenstein Center, which housed Donovan’s team and other media and politics researchers, tensions mounted. In her declaration, Donovan alleges that Laura Manley, the center’s executive director, tried to “sow discord” in a December 2022 meeting with an unnamed employee of hers. According to Donovan, who wasn’t there, Manley indicated that Donovan was looking out only for herself and planned to leave the team in the lurch. In Donovan’s telling, Manley encouraged the employee to report any problems with Donovan to human resources. Manley told me that Donovan “completely twists what happened.” Manley said that the vast majority of the conversation was about the employee’s future plans and that she tells all employees to consult HR with any questions they have about their affiliation with Harvard, regardless of the situation. Donovan claimed that an HR investigation deemed the incident “unprofessional conduct”; a Harvard spokesperson says this is inaccurate. Manley says she was never notified of any such investigation, as is required under Harvard policy. “This never happened, period,” she said.
On February 2, 2023, The Harvard Crimsonbroke the story that Donovan’s team would be folding and, with it, her position as the Shorenstein Center’s research director. A petition, started by a Kennedy School graduate and signed by more than 100 people, demanded her reinstatement and Elmendorf’s resignation. (He had already taken heat for denying a fellowship to a human-rights advocate, a decision that he ended up reversing.) Now the world knew the news that Donovan had been struggling to accept, in part by getting a new tattoo: a mermaid hovering over a book, sword held high. “I thought that that was a really beautiful metaphor for the protection of knowledge in society,” she told me. “And that this, you know, mermaid would sing this song to draw in all the bad dudes and then chop their fucking heads off.”
I feel like Joan wants us to talk publicly about how the dean has been pressured to get rid of her, but I don’t have any proof of that.
On the day of the Crimson story, Donovan suggested to her employees that the dean was not actually motivated by the policy he was citing — but, perhaps, by forces like Meta.
“The issue is that, ultimately, Shorenstein higher-ups felt like I had become dispensable, and thought that they could, in some ways, coordinate with the dean to get rid of this headache that I had become, because I was pretty insistent that we not take money from tech companies,” she said, according to a recording of a team meeting that I obtained. She alluded to wanting to found a nonprofit so “we’re not owned in any way.”
“Institutions like this can’t stand trailblazing,” she added. “The real story here is that our research has become so politicized — not political, but politicized — that people outside the university have been tapping on Doug’s shoulder, telling him, ‘You gotta get rid of them.’ This was a convenient excuse — it’s not even a good excuse.” Later, she said: “I would love for there to be a formal investigation into what Facebook has sent him about me. Or people from his Dean’s Council.” If lawyers or reporters contacted them, she told her employees, “just point them in the right direction.”
Some of them were conflicted from the start. “I feel like Joan wants us to talk publicly about how the dean has been pressured to get rid of her, but I don’t have any proof of that,” one texted another that day. “Not that I doubt it, but it’s not something I can actually speak to with any authority.”
ADVERTISEMENT
A little over a month after Schrage spoke up on Zoom, Harvard got $500 million for a new artificial-intelligence research center. It came from the Chan Zuckerberg Initiative, the organization founded by the Meta chief executive and his wife, Priscilla Chan, both Harvard alumni. Donovan’s lawyers point to this “significant financial influence” to assert that “how Facebook/Meta have operated in this case is no different to how foreign-intelligence services or organized criminal enterprises operate.”
But none of the money is going to the Kennedy School, and Donovan doesn’t offer a theory for how it might have affected her status there. “That gift just has nothing to do with me,” Elmendorf told me, “and it played no role whatsoever.” In general, “the allegations by Joan of unfair treatment and donor interference are false,” he said. (A spokesperson for the Chan Zuckerberg Initiative says it “had no involvement in Dr. Donovan’s departure from Harvard and was unaware of that development before public reporting on it.”)
A series of events in a suspicious order, a handful of well-connected people: This was what Donovan’s allegation boiled down to.
Similar logic was applied to the nuptials of Sheryl Sandberg, Meta’s former chief operating officer, a friend of Elmendorf’s since he was her undergraduate adviser. When he delivered the bad news to Donovan in August 2022, it was — in her words — “a mere four days” after he had attended Sandberg’s wedding on a Wyoming ranch (where, Donovan has also pointed out, he was photographed by People). The dean told me that he did not discuss the matter with Sandberg and that she “had no bearing whatsoever on the decisions regarding Joan Donovan.”
During that meeting with the dean, Donovan’s declaration says, Elmendorf told her that Harvard would “exercise its ownership of my book,” Meme Wars, because, unlike faculty, “all staff’s research was owned by the University.” Late last year, over dinner in Boston, she told me, “It is what it is: Someone can own my shit. I still know how to work a copy machine.” And in December, she tweeted, “The truth is H took everything from me,” including “my book,” and added, “I truly have nothing left to lose.”
But Harvard does not own the copyright to Meme Wars. By March of last year, the three authors and the provost had signed an agreement that “Harvard hereby irrevocably transfers and assigns to the Authors, in perpetuity and throughout the world, all of its right, title, and interest in the copyright” to Meme Wars, according to documents I obtained. (One exception: Harvard got a royalty-free license to use it “for Harvard’s research, educational, and other scholarly purposes.”) In Donovan’s declaration, the only reference to this agreement is a vague mention of the book being “settled.”
Elmendorf told me that transferring the copyright “seemed the fair thing to do.” And when I asked Donovan if it was misleading to not mention the agreement, she insisted that it was irrelevant because “to me, it is still very true that Harvard laid claim to my book.”
Meme Wars isn’t the only thing Donovan says Harvard took. She has made a series of accusations — at times ambiguous — that her ex-employer is “holding on to my intellectual property,” which the university broadly disputes. She recently asserted to me that Harvard has refused to negotiate with her lawyers over this issue since December (though she declined to put me in touch with them, saying that they do not want to talk to the media). According to Harvard, that is false. “We asked Joan a number of times before she left to tell us what IP she seeks,” a spokesperson told me. “Harvard’s counsel has welcomed conversations with her counsel since then, and has repeated our requests for Joan to identify what she is seeking. We have not heard back.”
Of these allegations, the most-detailed is that FBarchive — Harvard’s online trove of the leaked Facebook documents — was stolen from her by Latanya Sweeney, a professor of the practice of government and technology.
The first Black woman to earn a computer-science Ph.D. from the Massachusetts Institute of Technology, Sweeney went on to pioneer the field of data privacy and be the first to demonstrate that Google ads could be racially discriminatory. She was the faculty member who agreed for years to be the co-leader on Donovan’s grants from philanthropists, because Donovan was not allowed to lead that research alone. “I did it because otherwise, it just seemed like Joan wasn’t going to be able to bring in those grants or continue to operate at full capacity,” Sweeney told me. But “I did not like this idea for just so many reasons.” Because the research was overseen by Donovan alone, not by them together, Sweeney feared that their arrangement gave observers the wrong impression. “How do you say to a funder, ‘Yeah, that grant isn’t my grant, that’s Joan’s grant. My name is on it, but it’s not my grant,’” she said. “Instead, I just don’t ask them for a grant.”
That tension permeates their dueling histories of FBarchive. Donovan’s lawyers say that she was cut out of “her own project” because “Meta succeeded in co-opting Harvard”; Sweeney calls her former colleague’s version of events rife with “gross mischaracterizations and misstatements.”
Whose idea was it? On October 16, 2021, Donovan emailed Sweeney: “Would you want to talk with me and [another individual] about possibly making a searchable archive of these documents for the public interest?” She cites this message, plus a description that she wrote days later, as proof that the eventual website was her brainchild. She says that she got the files first and wanted the site to be like Genius, the song lyric-annotation platform. But Sweeney told me that she’d acquired her own, slightly different cache of the files before Donovan handed over hers. In Sweeney’s memory, Donovan understood the screenshots’ value and wanted to turn them into searchable text files, but “it was me who was saying no, there’s so much more that you can get from a full platform.”
Sweeney says that all the work of programming such a site fell to her and her team, the Public Interest Tech Lab, and it began on October 30, 2021, when she registered the domain FBArchive.org. Three days later, she started working on a slide deck describing a site where the content could be tagged and organized, according to a Google Sheets file, which she told me that she presented to people at Harvard.
From November 2021 through the summer of 2023, ranging from almost daily to weekly, she led or attended meetings about privacy, data, and technical development. Donovan was not invited because they fell outside the scope of what she could contribute, according to Sweeney: “She doesn’t build technology.” Donovan was, however, invited to 20 of the overall team meetings to discuss how to make the documents useful to researchers. She came to nine, according to Sweeney. She also showed up to about half of the 17 meetings she was invited to about the media rollout. Sweeney, who says that she personally wrote 10,000 lines of code and reviewed 20,000 screenshots, told me: “More than 100 people worked on this project. Every one of them will know me, very few of them will know Joan.” (Donovan told me, “I did attend as many meetings as I could make,” and said that she did work on her own time, which included making a list of research topics that the archive could help with.)
At one of the meetings Donovan missed, according to notes taken during it, an unnamed attendee asked what would happen “if Doug gets his feathers ruffled by Elliot,” referring to Elmendorf and Schrage. Lawyers for Donovan say this comment reflects the attendee’s view that the dean was “protective” of Meta. But Sweeney, who was there, told me that the person was alluding to Donovan’s concerns, not concerns of their own.
At another point before the archive’s launch, Sweeney proposed giving the dean a message to convey to Schrage and Sandberg: The site would benefit Facebook by getting “a thousand great minds to work on content-moderation approaches and other difficult issues.” She ran that phrasing by Donovan and Manley over email in April 2022. Donovan’s lawyers call the email a “concerning” turn in “what was supposed to be an independent research project,” saying it shows Sweeney and Manley helping the dean “placate the very company Dr. Donovan was researching.” But in the rest of the exchange, which Sweeney shared with me, Donovan herself helped write the note. She suggested, in a reply, explaining why the files were “imperative” to “understanding how Facebook adjudicates content-moderation decisions.” There is no evidence the message was ever sent to Elmendorf. Sweeney told me it was just one of several potential outreach strategies that both she and Donovan were brainstorming at the time to head off Facebook’s anticipated objections to the archive. (Donovan says she felt “uncomfortable” about the exchange, but “I was just trying to participate given that the decision had already been made.” Her lawyer told me the legal team had disclosed what it considered the relevant portion of the exchange.)
ADVERTISEMENT
By the time FBarchive was preparing to launch last year, the relationship between Donovan and Sweeney had deteriorated. In June 2023, the former accused the latter of removing her and her team’s names from the “about” page, which she called “gutting.” (Sweeney responded that the page was meant to reflect active contributors — the Technology and Social Change Project was by then being publicly phased out, she noted to me — and pointed out that Donovan and her team are mentioned on another page describing the site’s history.)
In October, FBarchive went live. Meta was allowed to review about 20,000 screenshots ahead of time and requested redactions to 816 of them. 161 got made, almost all of them names of employees who weren’t decision makers, and Sweeney’s lab had the final say.
Elmendorf told me that he isn’t aware of any other university that’s built a site like FBarchive. “I don’t understand an argument that the Kennedy School is somehow trying to protect Facebook when we have done that,” he said.
But Donovan says that the dean isn’t a reliable narrator, and that she wasn’t the only one with that opinion. “Dr. Sweeney once told me that she had attended a dinner at Ms. Sandberg’s home along with Dean Elmendorf,” her declaration says. “She was so struck by the closeness of their decades-long friendship that she confided in me her own concerns about Dean Elmendorf’s close, personal relationship with Ms. Sandberg and the potential conflict of interest with the work we were doing.”
Sweeney said that in April 2019, she did dine at Sandberg’s mansion in Silicon Valley, where a Harvard alumni-association representative had invited her to chat about her work with tech executives. But the dean told me that he wasn’t there. And Sweeney confirmed: “I’ve never been at a dinner with Sheryl and Doug. I’ve never been with the two of them at the same time ever.” She added, “I don’t know anything about Doug’s relationship with Sheryl Sandberg.” This story, she says, is a “lie.” (Donovan acknowledges that she “may have inferred” that Elmendorf was at the gathering, but she maintains that Sweeney expressed concern about the pair’s friendship.) Having now read Donovan’s whistle-blower disclosure, Sweeney told me, “I wouldn’t be able to trust her going forward at all.”
Five years ago, Donovan declared that it was high time for a beaver emoji. With its iconic buck teeth and webbed feet, it would charm lovers of semi-aquatic mammals and Canada. Plus, the sexting potential was strong. As Donovan said at the time, she and her then-partner “wanted lesbians all over the world to have proper emoji representation should they ever need it.”
It’s this side of Donovan — the funny, playful, disarming agitator — that her former employees loved. “She understands that we are people first with our own lives, our own struggles or triumphs, all these goals we’re trying to accomplish,” said Marya T. Mtshali, a former postdoctoral researcher for the Technology and Social Change Project, who praised Donovan for helping her organize a conference about misinformation in communities of color. “I initially looked at Joan as my colleague-slash-boss, and I left that experience with her being a really good friend.”
She’s punk rock in her soul, which is perhaps why she was never going to be able to stay at Harvard. Because Harvard is not punk rock.
Emily Dreyfuss
Alexei Abrahams, who built a web app that tracked right-wing extremists’ social-media activity, said that Donovan gave him a name for it (an internet “observatory”), encouraged him to write a book about it, and never tried to take credit. “Her management style or secret seemed to be to hire talented people and then get out of their way, or pave their way,” the former research fellow told me. Martin Rooke, another former fellow, praised the team as “truly a model of how scholars should work.” Dreyfuss, who edited the Media Manipulation Casebook, called Donovan “a wonderful boss” to whom she is deeply grateful for inviting her to co-write Meme Wars: “She didn’t have to do that at all.” She added: “I think she’s punk rock in her soul, which is perhaps why she was never going to be able to stay at Harvard. Because Harvard is not punk rock.”
But no one I spoke to could corroborate Donovan’s allegations about Meta. “At no point did I see any evidence of outside influence on the Facebook Archive project or any other research we were doing here,” Nancy Gibbs, the Shorenstein Center’s director, told me.
Some of Donovan’s former colleagues told me they flat-out do not believe her. And some told me that her rebellious charm could veer into something more volatile.
During her last two years at Harvard, Donovan was, by her own admission, distracted by her imploding personal life, namely her divorce from her partner of 17 years. She was also in and out of the hospital, sometimes for months at a time, with Covid-19 and other health problems. And from 2021 to 2022, management changes were afoot at the Shorenstein Center, which got a new executive director and finance director, and at the Kennedy School, where three people rotated through the role of financial-operations overseer. Staff members were trying to straighten out employees’ titles and duties by, for example, ensuring that research fellows weren’t doing full-time jobs, according to five people familiar with the situation. But Donovan read changes like these as disproportionate scrutiny of her work, and told me “it all stemmed from Doug getting upset with me.” (Elmendorf said he was not in a position to know about these kinds of personnel matters, which are handled by other faculty.)
Donovan’s disclosure mentions that during this time, Gibbs warned her of “repeated instances of excess spending,” but that she knew of only “one over-spending error.” In that case, documents show, leather jackets she gave to employees two years ago exceeded IRS gift limits to the point that they had to be taxed for them out of their paychecks. (Donovan says that she offered to pay the difference.) Emails show that other concerns about her spending were raised on other occasions. In 2022, she was warned in at least two instances of international travel expenses seeming unusually high. When she wanted to spend $13,800 on an event last year, an administrator told her that such big expenses almost certainly couldn’t be justified to funders and could no longer be approved. Donovan told me that she didn’t view these situations as “problems,” but as discussions that were resolved.
Some say Donovan was hard to pin down. “It was damn-near impossible to get a one-on-one meeting to talk about my research for more than a year,” said one former team member, who requested anonymity due to fear of retribution. From late 2022 onward, Donovan often canceled or failed to attend meetings, according to emails and four former colleagues. They recalled that when she did show up, she often complained about the dean and the Shorenstein directors instead of discussing work. (Donovan told me that “it’s very rare that I cancel a meeting” and that she was “meeting with team members consistently.” She also said that she made a point of not discussing “the blowback from Shorenstein” with them.)
“It truly was not a sound nonprofit institution,” Brandi Collins-Dexter, a former associate director of research, told me. “It felt to me like being the oldest kid in a neglected home and just trying to make sure everybody got dinner every night.”
Collins-Dexter had met Donovan after the 2016 election, when she was working with the advocacy group Color of Change. She went on to run a high-profile campaign pressuring Facebook to audit its civil-rights policies on hate speech and voter suppression, and to help organize a boycott with hundreds of advertisers. When she started feeling burned out, Donovan invited her to do a fellowship, which she used to begin researching a book about Black voters in America. Then in the summer of 2022, Donovan asked her to help run the team. “I felt like she had been a lifeline and safety net, and I felt so loyal to her,” Collins-Dexter told me. But she eventually came to question her boss’s character.
ADVERTISEMENT
In the fall of 2022, Donovan and Collins-Dexter were texting about their misgivings with the “disinformation creep” paper that had appeared in the Kennedy School’s journal. The fallout from the retraction was still not over, now nearly a year later. That October, Donovan heard from the sociologist who accused her of engineering the retraction and blamed her for, among other things, having “seeded distrust” with groups funding Mutale Nkonde, the lead researcher and then a master’s student at Columbia University. (Donovan denied to me that she was responsible for any problems Nkonde had with funders. Nkonde declined to comment.)
Around that time, Collins-Dexter was preparing to give a talk to funders about her work. Donovan suggested that she mention Nkonde’s retracted paper as an example of “how funders eager to support black women aren’t vetting the field,” according to texts Collins-Dexter shared with me. (She initially agreed, but later changed her mind, she told me.)
In April 2023, Nkonde herself texted Donovan seemingly by accident. “Joan has no money to take [to] her new institution and is demanding tenure,” she wrote, adding a string of laughing emojis. “So far everyone has said no.”
Donovan mulled whether to hit back in ways that she, as a media-manipulation expert, intimately understood. “What do I do?” she asked Collins-Dexter, sharing the screenshot with her. “I want to post it to twitter.”
“That could backfire,” her colleague cautioned.
I felt like I was being used as a shield and a weapon in service of Joan.
Brandi Collins-Dexter
But later that day, Donovan reported: “I made good gains with academics that follow me on Instagram after I posted the screenshots. Lots of them were wondering what had really happened.”
Donovan then raised the idea of publicizing details of the dispute — anonymously. She asked Collins-Dexter: “Do you have any sock puppet accounts?” A sock puppet, as defined by Donovan’s own team, is a “false online identity typically created by a person or group in order to promote a specific narrative or opinion, sow division, or circumvent a previous account ban.”
Again, Collins-Dexter refused to entertain the idea.
When asked, Donovan told me that she never actually posted the screenshots to Instagram. She also claimed that she was “joking” about sock puppets. “I certainly have never tried to stir the pot on someone in this way,” she told me.
But Collins-Dexter told me that in hindsight, as a Black woman, “I felt like I was being used as a shield and a weapon in service of Joan — Joan’s brand and Joan’s public perception and accusations against her by other Black women.” (Donovan said that the notion that she was “playing some kind of identity politics game” is “disingenuous,” adding that half of the full-time staff she’s ever hired have been people of color and a third have been women of color.)
Collins-Dexter now wonders if there was any truth to the accusation that Donovan had orchestrated a “whisper campaign.” Criticizing research was one thing, she told me, but it was “another to try to get people defunded, ruin their careers or potentially be subjected to online harassment.” It was not lost on her that Donovan’s work was, theoretically, concerned with preventing the latter.
By the spring of 2023, the Technology and Social Change Project was entering what would be its fraught final stretch — and its leader was openly alleging to administrators that Harvard was engaging in “a cover up of wasted funding.”
During this period, the dean had instructed Donovan to not take on new projects, including large conferences. She told higher-ups in an email, quoted in her declaration, that these rules were “preventing Harvard from meeting contractual obligations” to donors. Most egregiously, she alleged, she was being blocked from hosting a required conference on the theme of a public-interest internet. Given the circumstances, she “must be clear with all funders of these risks,” she declared in an email to Gibbs, the Shorenstein Center’s director. (Gibbs told her that, as a matter of policy, all communications with funders had to be in writing and coordinated with the university, and that she shouldn’t tell them to stop their grants or that Harvard couldn’t fulfill them. Harvard says it’s up to the school to approve a grant transfer request, and only under certain conditions.)
All the while, Donovan’s declaration says, she was “terrified for my staff, who were in need of contract renewals, stable employment and healthcare.”
But throughout last spring, according to texts, emails, and people who worked with her, Donovan was talking to her donors about ending their funding commitments — even though doing so would put her staff out of work even faster than anticipated. (Donovan denied approaching funders about pulling their money, saying, “All of my funders contacted me.”)
After the Crimson reported in February 2023 that their team would be eliminated, Collins-Dexter told me, Donovan asked her to arrange calls and meetings with some of their donors to discuss its future. She organized one such dinner with employees from two foundations, at a French restaurant in New York City, after Donovan initiated contact with them, emails show. Based on conversations beforehand, Collins-Dexter says, she assumed that her boss was going to reassure them that their money was safe until June 2024, the end date publicized in the news. That would’ve been just fine with Collins-Dexter and some of her colleagues, who were telling Donovan that they wanted to stay on the payroll as long as possible, according to texts from the time.
ADVERTISEMENT
Instead, Collins-Dexter recalls, Donovan told funders to reroute their grants into a nonprofit that she wanted to start, to continue her research post-Harvard. “I was really shocked when I heard Joan start asking funders to pull their funding,” Collins-Dexter told me. She believes that after she tried to push back, she stopped getting invited.
Two major donors, the William and Flora Hewlett Foundation and the John D. and Catherine T. MacArthur Foundation, had together given or planned to give about $1.8 million to the team. On March 3, 2023, when Donovan texted a teammate about a new funder interested in the nonprofit, she added, “We will get it started with that Hewlett $$ and I’m seeing if MacArthur will divert funds.” She made it official in an internal email days later: “I spoke to MacArthur and Hewlett and they are both going to end their grants early.” By the end of the month, MacArthur was canceling a planned $150,000 payment and Hewlett was canceling an expected $300,000, according to an internal budget document. (Donovan says that employees from Hewlett and MacArthur volunteered, in conversations with her, to back out of their pledges. A Hewlett spokesperson says the foundation, which now funds Donovan’s nonprofit, consulted both her and Harvard in ending its grant early, since the team that “the grant was intended to support would no longer be happening there.” A MacArthur spokesperson says that it was notified by an individual at the Shorenstein Center — whom it declined to name — that Harvard would not finalize the grant.)
MacArthur, Hewlett, and a third sponsor, the Ford Foundation, were helping bankroll Donovan and 18 of her employees, according to the internal budget document. In a finance meeting on March 24, 2023, Donovan told a handful of staffers that she wanted to use $80,000 from Ford — money that would normally go to staff salaries — to hold the public-interest-internet conference that she claimed was mandatory, according to meeting notes I obtained. If the event didn’t happen, she wanted to return any unused funding to the foundation.
But a research-funding administrator at Harvard told her she wasn’t actually contractually obligated to host the conference, according to the notes. Donovan appeared determined to hold it anyway. “I feel for the rebudget that people might lose their jobs,” she told the group, according to the notes, “but I can’t compromise the integrity of the research.” Later that spring, a Ford employee would affirm in an email to Harvard staff that the grant didn’t have to be used for an event or any other specific project. Last week, a Ford spokesperson confirmed to me that the grant had no restrictions.
As of last month, Donovan still believed that the conference had been a requirement. “Anybody at Shorenstein that’s telling me what to do,” she told me, “they don’t know the contracts like I do.”
The event ultimately wasn’t held. Ford did not withdraw its funding, according to the spokesperson, and nobody at the foundation spoke with anyone at the Shorenstein Center about doing so.
How I see Joan more than anything is an opportunist who will stop at nothing to put herself in an optimal position.
Brandi Collins-Dexter
For all of Donovan’s talk of self-sacrifice, Collins-Dexter felt betrayed by her actions. “I think all of us would have been prepared to go hard in the paint over the next eight months for this work and this team if we felt like Joan was prepared to fight for that,” she told me. “Instead, she did everything that she could to sabotage it and undermine it.” Donovan disputes this characterization, saying that her employees were not privy to her conversations with funders.
In Collins-Dexter’s eyes, her friendship with Donovan ended in September 2023, at a conference in New York City. According to her, Donovan told her, crying, that Harvard had stolen “our money” and that it was going to take the copyright to Collins-Dexter’s book. Collins-Dexter, who felt confused and angry at being confronted in public, strongly doubted the first claim and knew, as a lawyer by training, that the second wasn’t legally possible. Later, she would wonder if the episode was a test: Would she corroborate Donovan’s allegations once public? (Donovan told me that was not her intent.)
To Collins-Dexter, the theory that Meta intimidated Harvard into disbanding the team makes no sense: She was hired despite having testified before Congress about Facebook’s disinformation problem, and she and a half-dozen former colleagues remain at Harvard, in her case on a fellowship.
“How I see Joan more than anything is an opportunist who will stop at nothing to put herself in an optimal position,” Collins-Dexter told me. “Anyone who’s seen as a threat to that is somebody to be targeted, displaced, and she doesn’t care who those people are.”
Donovan says that’s not true. “It breaks my heart, it really does.”
Donovan’s last day at Harvard was August 31, 2023 — 10 months earlier than the end date Elmendorf had originally given her. There are, once again, competing narratives as to why. In her telling, she “advocated for full extensions of the entire team through June 2024,” but HR wouldn’t grant all of them, even though it had “millions in funding.”
A Harvard spokesperson says that administrators initially thought the team’s work and funding could last through this month if needed, but by mid-2023, “a substantial amount” of the work was complete and many staff members had found new roles. Also, a lot of the money was gone. Most of the team’s grants were scheduled to run out by the end of August, according to an internal budget document, at which time there was $1.7 million, as an HR employee told Donovan by email. (The Harvard spokesperson confirmed that there were not millions remaining by the time Donovan left.) There had, at one point, been an expected grant from the Hewlett Foundation that could have helped carry the staff until September 2024. But it had been canceled.
Harvard says it offered to let Donovan continue as a part-time adjunct lecturer, but she chose not to. In August, she announced that she was joining Boston University — the place offering “the most academic freedom” — as a tenure-track assistant professor of journalism and emerging media studies. When I visited that fall, she was off to an ambitious start. In a packed lecture hall, she gave a talk on “memetic warfare” in politics. She told me of her plans to archive the online posts of politicians around the world. We picked up boxes of T-shirts for her nonprofit, which by then had a name: the Critical Internet Studies Institute. Later, I asked her dean, Mariette DiChristina, what Donovan had told her about why she’d left Harvard. She declined to comment.
Five months after we met in Boston, I confronted Donovan with the findings in this story. Over a call that lasted nearly three-and-a-half hours, she responded to each claim against her, at times in tears.
ADVERTISEMENT
I ended with the question that had become impossible to ignore: Had she — the famed misinformation researcher — been spreading misinformation?
She was silent at first. “I do stand by what’s in the declaration,” she finally said, “and what I’ve said to you and how I’ve presented what happened to me. I didn’t make up anything. If I had, it would be easy to disprove.” Her voice broke. “I’m merely one woman in the world,” she went on, “and that’s it. And to go up against corporations like Harvard and Facebook is very scary. But that is my truth. That is what I know to be true — that I was on top of it one day, and I was under it the next.”
We hung up. Starting that night and over the following four days, Donovan texted me 88 times. She forwarded me emails, texts, screenshots, voice mails. Each, she insisted, showed the facts to be on her side.
Her messages grew longer, darker. Unprompted, she began to refer to shadowy forces working against her. “There are people who do want me dead,” she wrote. “This line of work is wild and I don’t wish it on anyone. I am looking into changing my legal name so that it’s difficult to hack or find out where I live, by socially engineering access to my banking or property records.” She wrote that she was afraid of being assaulted, of being murdered.
“Harvard and Meta want negative press about me,” she wrote, “and that could get me killed, especially if one person gets it in their head that I’m illegitimate in some way.”
Two powerful institutions were out to discredit her, and my article was just more proof.
“This shows me,” she wrote, “that Harvard and Meta chose you.”
Stephanie M. Lee is a senior writer at The Chronicle covering research and society. Follow her on Twitter at @stephaniemlee, or email her at stephanie.lee@chronicle.com.