After Trump’s Election, NSF-Funded Censorship Tools Go Missing

The National Science Foundation (NSF) claims to be the place “where discoveries begin” — but it funded programs where free speech ends. When The Federalist asked if it still funds anti-“misinformation” tools, a representative simply pointed to its public grants database, which lists censorship projects funded during the Biden administration.

The public-interest law firm Alliance Defending Freedom sent NSF records requests in February to uncover any coordination between the federal agency and internet communications monopolies such as Google and Facebook.

NSF funds one-quarter of all federally funded academic fundamental research projects at U.S. higher education institutions, and funds technology development at approximately 400 “small businesses” each year. It has an annual budget of $8.5 billion. It also funds programs that exclude recipients based on race, according to an NSF factsheet.

The NSF’s “Convergence Accelerator,” which funds special research projects, launched a “cohort track” in 2021 for “Trust & Authenticity in Communication Systems.” Recipients developed software to control online speech labeled “misinformation” about politically sensitive topics including Covid-19 treatments and election integrity.

Twelve “teams” were selected for “phase 1” of the project, according to the House Judiciary Committee, and six were selected for “phase 2” funding at $5 million each. In total, the NSF awarded $39 million total to projects in this “track” of the program.

The NSF-funded projects include the Analysis and Response Toolkit for Trust (ARTT), Co-Designing for Trust, Co:Cast, Co-Insights, CommuniTies, CourseCorrect, Expert Voices Together, Search Lit, TrustFinder, and WiseDex. Projects had ties to infamous pro-censorship organizations including Google, Meta, Snopes, Wikimedia, the World Economic Forum, and the World Health Organization.

The Federalist asked NSF how much total funding each project received from the agency, and if there have been any “cohorts dedicated to developing anti-misinformation projects since ‘track F.’” An NSF spokesman replied that, “in recent years,” Congress asked NSF to “identify and address issues of safety, ethics and adversarial influence online” through funding bills and the IOGAN Act. So the Convergence Accelerator “initiated Track F.”

“This program has not made an award since 2021 and will not be making any awards in the future,” the representative said. “NSF invests in research, innovation, and workforce development that accelerates the development, testing, and understanding of technology. NSF plays no role in content policies nor content regulations.”

The Federalist again asked “how much total NSF funding each project received” and “whether NSF helped develop other anti-misinformation projects aside from Track F.” The spokesman simply pointed to the agency’s “award search page,” and said, “[o]ther than that, I don’t have anything else to add.” A search for active NSF awards regarding “misinformation” yields more than 100 results.

ADF also filed public records requests for documents regarding the NSF Convergence Accelerator, as The Federalist previously reported. Mathew Hoffman, legal counsel for ADF’s Center for Free Speech, said at the time the group was investigating “where our tax dollars are being spent to fund censorship” — and “if anyone’s rights have been violated by the censorship-industrial complex, that litigation will certainly be an option.”

Manipulating Public Discourse

The Analysis and Response Toolkit for Trust’s (ARTT) “main tool” was created to be a “Web-based software assistant that provides a framework of possible responses” in controversial discussions. The NSF awarded ARTT nearly $750,000 in 2021. ARTT says it also received $5 million in startup funding. In a video online, Jim Chen with the University of Washington (UW) showed how the chatbot responds differently as a Republican or Democrat.

This year, ARTT “spun off” as a nonprofit, Discourse Labs. Its “Principals and Board” include Connie Moon Sehat, a former senior fellow for the World Economic Forum.

“Discourse Labs is a nonprofit with the mission of developing tools and resources for productive public dialogue, which includes the Analysis and Response Toolkit for Trust (ARTT),” a representative for the group told The Federalist. “ARTT received funding from the National Science Foundation from 2021 to 2024.” The nonprofit has not yet posted any public tax documents revealing new funding sources.

Faculty from multiple universities spearheaded Expert Voices Together, a “rapid-response system” for “journalists and researchers.” The NSF Convergence Accelerator awarded EVT nearly $750,000 in 2021 and $5 million in 2022, and the group became “housed” at the leftist group Right To Be this year. A representative for Right To Be told The Federalist EVT “remains under the direction of George Washington University,” and to “direct inquiries there.” GWU did not respond to repeated requests to comment.

While those federally funded tools may still be active, others claim they’ve been discontinued. The University of Michigan (UM) helped develop WiseDex for “Fast, Comprehensive Misinformation Enforcement,” helping online platforms push “difficult judgments to someone outside the company … by externalizing the difficult responsibility of censorship,” per notes from UM’s first pitch to the NSF, according to the House Judiciary Committee.

WiseDex planned to work with online speech control groups Meedan and the infamous Election Integrity Partnership. The NSF Convergence Accelerator awarded WiseDex $750,000 in 2021, but its phase two proposal was “not funded.” Paul Resnick, director of UM’s Center for Social Media Responsibility, told The Federalist WiseDex “is not currently active.”

The University of Washington helped develop Co-Designing for Trust, which claims to quell speech deemed “misinformation” in “(BIPOC) and rural communities.” Jason Young, a UW employee, called it a “Black Snopes” in a video. Spokesman Victor Balta told The Federalist it “improve[s] educational resources related to information literacy,” like a project to help rural library patrons “avoid being scammed on Craigslist.” He said the project is “winding down as the grant-funding period approaches its conclusion.” The NSF awarded Co-Designing for Trust $750,000 in 2021 and $5 million in 2022.

UW also developed TrustFinder, which was supposed to help “researchers to reduce the negative impacts on their work of misinformation.” Balta said the project “did not advance beyond the initial planning phase in 2021 and is no longer active.” The NSF Convergence Accelerator awarded TrustFinder $5 million in 2022.

Meedan developed Co-Insights — initially named Fact Champ — to “scan millions of posts” and fight “misinformation targeting Asian-American and Pacific Islander (AAPI) communities.” NSF awarded it nearly $750,000 in 2021 and an additional $5 million in 2022. A Meedan representative told The Federalist the project “studied and published peer-reviewed research on content targeting communities online. …” The representative claimed Meedan “strongly support[s] free speech” so its “tools were opt-in and only provided additional information,” and the project “ran from 2021–2024 and has now concluded.”

Ohio State University also developed the Crisis Observatory: Credible Actionable Sensemaking Tools, an “AI-based system” that “separates credible from less credible information.” The NSF awarded it nearly $750,000 in 2021. Benjamin Johnson, chief spokesperson for OSU, said it was “disaster response research to aid first responders,” and researchers studied how AI could assist them and “other decision makers analyze large amounts of information in a rapidly changing emergency situation like a natural disaster or public health crisis.”

All the discontinued censorship tools mean taxpayers paid tens of millions of dollars for nothing at all — except keeping numerous censorship jobs temporarily afloat.

Logan Washburn is a staff writer covering election integrity. He is a spring 2025 fellow of The College Fix. He graduated from Hillsdale College, served as Christopher Rufo’s editorial assistant, and has bylines in The Wall Street Journal, The Tennessean, and The Daily Caller. Logan is from Central Oregon but now lives in rural Michigan.