Background on the SIO’s Projects on Social Media

Background on the SIO’s Projects on Social Media

stanford dish at sunset Linda A. Cicero

Various inaccurate and misleading claims have been made in the media and in congressional testimony regarding the Stanford Internet Observatory’s projects to analyze rumors and narratives on social media relating to U.S. elections and the coronavirus. As explained in the statement issued by our partners at the University of Washington, it is difficult to rebut all of these inaccurate claims without repeating the falsehoods and contributing to their further spread. Nevertheless, the SIO believes it is important to respond to inaccurate statements about its work and to correct the public record. This fact sheet provides background on the SIO’s activities and corrects a number of false allegations.

  1. The Stanford Internet Observatory, founded in June 2019, is a non-partisan, cross-disciplinary program of research, teaching and policy engagement for the study of abuse in current information technologies, focusing on social media. The SIO studies and publishes about disinformation and state influence operations and conducts and publishes research regarding child internet safety, online platforms’ policies and practices toward self-harm, and privacy-protecting technologies.

  2. The SIO is one of four organizations that convened the Election Integrity Partnership (EIP) and Virality Project (VP). The EIP and VP, both founded in 2020, are non-partisan research coalitions that operate in an open, transparent, and public manner, publishing blog posts, weekly updates, briefing videos, academic papers, and voluminous investigative reports relating to election and vaccine misinformation, disinformation, and propaganda.

  3. The EIP’s goal was and continues to be to research and analyze attempts to prevent or deter people from voting, as well as efforts to delegitimize election results. The EIP did not study online discussions of specific candidates, parties, or political topics in the 2020 or 2022 election cycles. The EIP did not make recommendations to social media or take any other actions regarding content about the Hunter Biden laptop story.

  4. The VP was established to research viral narratives on social media related to COVID vaccines in four areas: “(1) safety, (2) efficacy and necessity, (3) development and distribution, and (4) conspiracy theory[.]” Its primary goal was to facilitate awareness for public health officials and medical professionals seeking to communicate accurate information to the public.

  5. The EIP and VP provided public factual findings to multiple entities, including government agencies and social media platforms, but had no control over content moderation, censorship, or labeling posts. Emails released in the Twitter Files demonstrate that Twitter’s staff examined any reports sent to them to see if the content was violative of their policies and took no action in cases where they felt that Twitter’s existing policies were not violated.


Was Stanford Internet Observatory’s Election Integrity Partnership created in response to public criticism of DHS’s “Disinformation Governance Board” and to substitute for the work of the government?

No. The EIP was created in the summer of 2020, long before the announcement and media criticism of the Disinformation Governance Board in April 2022.

Does the SIO or EIP receive funding from the federal government?

As part of Stanford University, the SIO receives gift and grant funding to support its work. In 2021, the SIO received a five-year grant from the National Science Foundation, an independent government agency, awarding a total of $748,437 over a five-year period to support research into the spread of misinformation on the internet during real-time events. SIO applied for and received the grant after the 2020 election. None of the NSF funds, or any other government funding, was used to study the 2020 election or to support the Virality Project. The NSF is the SIO’s sole source of government funding.  

Is it true that the EIP censored 22 million tweets and labeled them as “misinformation”?

No, the EIP did not censor any tweets or label any tweets as “misinformation.” EIP has no ability to remove or label tweets or other posts, and content moderation decisions are independently made by social media platforms. As part of its non-partisan research relating to the 2020 U.S. presidential election, EIP analyzed 22 million tweets that contained keywords or URLs relevant to EIP’s scope of work. EIP identified 2,890 unique tweet URLs in potential violation of Twitter's stated policies.  EIP provided its factual analysis to the relevant platforms, which were then responsible for each platform’s own content moderation decisions. The EIP informed Twitter and other social media platforms when certain social media posts violated each platform’s own policies; EIP did not make recommendations to the platforms about what actions they should take.

Did the EIP “target” or discriminate against conservative social media accounts or content or seek to promote liberal accounts or content?

No. EIP is a non-partisan coalition dedicated to the identification and analysis of online content that suppresses voting, reduces participation, confuses voters about election processes, or delegitimizes election results without evidence. Based on an analysis of an expansive dataset and without targeting any specific accounts of politically affiliated content, EIP’s research determined that accounts that supported President Trump’s inaccurate assertions around the election included more false statements than other accounts. Through the same analysis, EIP also found and published or flagged examples of false statements and rumors made by accounts linked to left-leaning groups and foreign actors. EIP informed social media platforms of examples of violative content by progressive groups as well as conservative groups.    

Did EIP receive direct requests from the Department of Homeland Security’s Cybersecurity and Infrastructure Security Agency (CISA) to eliminate or censor tweets?

No. In its non-partisan research during the 2020 and 2022 elections, EIP analyzed reports of potentially false information received from a broad array of sources, including state and local election officials. These reports were channeled through the Election Infrastructure Information Sharing & Analysis Center (EI-ISAC). The EI-ISAC did not ask the EIP to censor or eliminate social media posts. EIP’s role was limited to the analysis of potentially inaccurate information that had been reported to EI-ISAC by state and local election officials. As noted above, the EIP had no ability to censor or eliminate social media posts; it simply identified potentially policy-violative posts to social media platforms.

Did the SIO’s Virality Project censor social media content regarding coronavirus vaccine side-effects?

No. The VP did not censor or ask social media platforms to remove any social media content regarding coronavirus vaccine side effects. Theories stating otherwise are inaccurate and based on distortions of email exchanges in the Twitter Files. The Project’s engagement with government agencies at the local, state, or federal level consisted of factual briefings about commentary about the vaccine circulating on social media.

The VP’s work centered on identification and analysis of social media commentary relating to the COVID-19 vaccine, including emerging rumors about the vaccine where the truth of the issue discussed could not yet be determined. The VP provided public information about observed social media trends that could be used by social media platforms and public health communicators to inform their responses and further public dialogue. Rather than attempting to censor speech, the VP’s goal was to share its analysis of social media trends so that social media platforms and public health officials were prepared to respond to widely shared narratives. In its work, the Project identified several categories of allegations on Twitter relating to coronavirus vaccines, and asked platforms, including Twitter, which categories were of interest to them. Decisions to remove or flag tweets were made by Twitter. 

More information regarding the VP’s research regarding vaccine-related narratives and engagement with social media platforms and the government is available through the VP weekly briefings, which were posted publicly throughout the Project, and in the VP’s final report.