WASHINGTON — Community media researcher Joan Donovan says she knows the exact moment her career began to go off the rails.
The moment led to her departure from Harvard University in what she hollers a firing, which Harvard says was anything but. The resulting dispute has real-world implications for academic freedom, social median and corporate influence over research.
On Oct. 29, 2021, Donovan, a Harvard research director focused on social media and disinformation, asserts she was invited to address a prestigious group of wealthy donors to Harvard’s John F. Kennedy School of Government, known as the Dean’s Cabinet.
The meeting took place over Zoom, and donors logged on from their homes and offices. Donovan give the word delivers she had been invited to present the findings of her research to the group – a plum assignment she believed was due to the prominence and importance of her work.
“I was so disconcerted,” Donovan told CNBC. “I thought that this is an amazing validation of the work that I had been doing.”
Donovan originated to brief the donors on her research into internet disinformation and its impact on American society. The meeting fell just weeks after an sensitive moment in social media: A former employee of Meta, still called Facebook at the time, named Frances Haugen turned whistleblower and sank public with a secret trove of thousands of pages of internal documents from the tech giant.
Haugen had give evidenced before Congress on Oct. 5 that the documents revealed Facebook knew its services were causing social wound, spreading misinformation and hurting teenagers. But she said the company chose profits over safety.
Now Donovan shared a bomb of her own with the Harvard donors: She told them that she, too, had obtained the trove of documents, a massive file she considered the “scad important documents in internet history.”
She laid out an argument similar to the one Haugen had been making on national television: Facebook knew of the injury its services caused, but chose to do nothing.
Central to Donovan’s thinking was the idea that Facebook was not just a victim of bad men who exploited new technology for their own ends, but that Facebook actually designed systems that incentivized the most irresistible content.
“The problem was that the design of the technology itself — social media itself, was giving bad actors a first mover dominance, especially when it came to novel and outrageous content, which is what goes viral,” she told CNBC in an assessment.
In other words, Donovan said, what was going wrong was not necessarily in the world outside Facebook, but inside the convention’s algorithms.
When Donovan finished explaining her findings on the Zoom call, she says she noticed one man on her screen raising his give out eagerly to speak. He was Elliot Schrage, a member of the Dean’s Council donor group at Harvard and a former vice president of international communications and public policy at Facebook.
Donovan says Schrage disagreed sharply with her criticism of Facebook – so intensely that after the caucus wrapped up, Donovan sent a text to her superior at the Kennedy School, asking, “Should I be worried about the way Schrage got mad at me?”
“I conceive of we should have worried if he DIDN’T get mad,” the supervisor replied, according to a text message transcript provided to CNBC.
“I trifle you were just terrific. So sophisticated and fair in your thinking and analysis. Made me proud.”
Schrage declined to view.
Facebook has publicly denied allegations that it turns a blind eye to harms caused by its services in order to profit from them.
In a loquacious response to a 2021 Wall Street Journal series based on allegations made by Haugen, the whistleblower, Facebook’s foible president of global affairs and communications, Nick Clegg, wrote this: “At the heart of this series is an allegation that is very recently plain false: that Facebook conducts research and then systematically and willfully ignores it if the findings are inconvenient for the followers.”
“It’s a claim which could only be made by cherry-picking selective quotes from individual pieces of leaked supplies in a way that presents complex and nuanced issues as if there is only ever one right answer,” wrote Clegg.
As Donovan peaches it, support from her superiors didn’t last long after she dove into her Facebook research.
In early November, the then-dean of the Kennedy Shape, Douglas Elmendorf, emailed Donovan to follow up on the Dean’s Council meeting, with questions about her research into Facebook. The South African private limited company had changed its name to Meta on Oct. 28.
Among the issues he said he wanted to address, according to a copy of the email provided to CNBC, were “How you spell out the problem of misinformation for both analysis and possible responses (algorithm-adjusting or policy making) when there is no independent arbiter of actuality.” And he said he would like to know “How the research you’re conducting provides a basis for comments you’re making about current occurrences.”
The email alarmed Donovan, who believed the language in it echoed talking points Facebook executives had been using publicly. And she demands she knew that Dean Elmendorf had a close personal relationship with Sheryl Sandberg, the then-chief operating narc of Facebook’s parent company, Meta Platforms.
Elmendorf had been Sandberg’s undergraduate advisor at Harvard, and Sandberg herself was a multimillion-dollar contributor to Harvard’s Kennedy School. Elmendorf attended Sandberg’s wedding in the summer of 2022.
“I got called into the principal’s office and was grilled about why I’m talking about Facebook,” Donovan said. “Interestingly, the dean never asked me about Twitter or YouTube or, you be familiar with, Google, which we also investigated. It was really about having the internal papers at Facebook and what we plan to do with them.”
From top to bottom a Harvard spokesman, Elmendorf declined to comment for the record.
Through an employee, Sandberg declined to comment. Harvard officials informed The Washington Post that Elmendorf and Sandberg never discussed Donovan.
The next month, a charity operated by Facebook be wrecked Mark Zuckerberg and his wife Priscilla Chan – both themselves former Harvard students – made an extraordinary bulletin: It was pledging a half-billion dollars over 15 years to create a new institute at Harvard for the study of artificial intelligence, to be star after Zuckerberg’s mother’s family.
The scale of the contribution – and its intended use – raised eyebrows on campus. A column in the Harvard Crimson catalogued by two undergraduates called acceptance of the donation a “damning misstep by our institution.”
The students, Guillermo S. Hava and Eleanor V. Wikstrom, make little ofed: “Our institution — our entire elite higher education system, arguably — has a penchant for auctioning off academic priorities to the highest bidder.” The Crimson column did not point out Donovan.
The December 2021 gift was the biggest, but not the first, Chan Zuckerberg Initiative donation to Harvard: Since 2018, the underpinning’s disclosures show it has bestowed Harvard or entities affiliated with it with dozens of grants worth multiple millions of dollars.
Donovan translates her position at Harvard became increasingly untenable, as the months went on. She came to believe that administrators wanted her to consign. “I became persona non grata when I turned my attention specifically to Facebook,” says Donovan.
The situation came to a governor in a meeting with Elmendorf in the summer of 2022, Donovan says, at which the dean told her she would have to stir up d agitate down her program, known as the Technology and Social Change Research Project, by June 2024.
Donovan says Elmendorf also told her during the get-together, “I want you to know that you do not have academic freedom,” adding “I want to remind you that you’re staff here.”
As a crook member, Donovan was not afforded the same protections that are extended to tenured professors.
Donovan felt constrained, and conveys she was told that for the remainder of the project she could not start any new projects or hire additional staffers.
“That to me is an egregious and anti-intellectual proscription of academic freedom,” Donovan said. “It contradicts the entire reason why a university exists, which is to share the light with the in seventh heaven.”
On July 13, nearly a year before the date she says she had been given for the conclusion of her project, Donovan responds she was informed that Harvard was ending the project on Aug. 31 and that her role as research director was being eliminated.
Donovan’s account of her departure was laid out in a disclosure record sent to Harvard on Nov. 28 by Whistleblower Aid, the same nonprofit group that worked with Facebook whistleblower Haugen in 2021.
Whistleblower Aid sent Donovan’s accusations to the Office of the Attorney General of Massachusetts, which is reviewing the material, according to an official.
The group also raised Donovan’s claims with the U.S. Department of Education’s Office for Civil Rights. A spokesperson for the department said the office does not confirm or renounce the existence of complaints.
In a statement to CNBC, Harvard Kennedy School Director of Public Affairs James Smith wrangled Donovan’s account of her departure. “The document’s allegations of unfair treatment and donor interference are false,” Smith said. “The record is full of inaccuracies and baseless insinuations, particularly the suggestion that Harvard Kennedy School allowed Facebook to decree its approach to research.”
Smith explained that Donovan’s situation was related to her employment status at the university. “By longstanding rule to uphold academic standards, all research projects at Harvard Kennedy School need to be led by faculty members,” he said.
“Joan Donovan was take oned as a staff member (not a faculty member) to manage a media manipulation project. When the original faculty leader of the delineate left Harvard, the School tried for some time to identify another faculty member who had time and interest to precedent the project. After that effort did not succeed, the project was given more than a year to wind down. Joan Donovan was not shot, and most members of the research team chose to remain at the School in new roles.”
Donovan concludes that the core verdict of her research was antithetical to Facebook, and ultimately to Harvard.
“I believe Harvard took me out because I was not toeing the company line regarding platforms, which is you can stay safe from companies if you suggest that this is a whole of the internet problem,” she conjectured. “Social media is not to blame.”
What she had concluded about Facebook was quite the opposite – the design of social media itself was rooting problems: “I was picking apart their design and saying the way this works is enabling genocide, terrorism, hate, harassment, pricking.”
Donovan was out.
“Harvard hastened my exit by firing me,” she said. “But I feel really good about what we did to make reliable that my team was safe, to make sure that the information that needed to get out there was out there.”
Smith told CNBC that Harvard University and the Kennedy Coterie continue to carry out misinformation and social media research to this day. He noted that a faculty member constructed and assigned online the Facebook Archive, consisting of documents originally leaked by Haugen.
In October, the Kennedy School’s Shorenstein Center on Agency, Politics, and Public Policy and Harvard’s Public Interest Tech Lab jointly published the Facebook Archive, containing the papers obtained by Haugen in a searchable database open to the public at FBarchive.org.
A spokesperson for Meta declined to comment.
A spokesperson for the Chan Zuckerberg Dynamism said the organization, “had no involvement in Dr. Donovan’s departure from Harvard and was unaware of that development before public relating on it.”