Meta Prepares Algorithm Change to Reduce Reach of Federal-Flagged Posts, Documents Show

The Algorithm Change

Meta is preparing to roll out a new content ranking adjustment that would reduce the distribution of posts flagged by federal agencies, according to internal documents reviewed by The Alamo Post and two employees at the company familiar with the system. The change, referred to internally as Project ShieldRank, is scheduled for a phased launch beginning Feb. 14 and would apply to Facebook and Instagram content flagged by the Department of Homeland Security, the Federal Bureau of Investigation, and the Department of Health and Human Services, the employees said.

A 34-page internal memo dated Jan. 27, titled "Content Integrity Quarterly Review: First Quarter 2026," describes Project ShieldRank as a "demotion signal" that would lower the ranking of posts identified by partner agencies as containing election misinformation, public health misinformation, or foreign influence content, according to one of the employees. The memo was distributed to roughly 120 employees in Meta's Trust and Safety division, the employee said.

The system would not remove flagged content outright but would reduce its distribution in users' feeds by an estimated 60 to 80 percent, according to the second employee, who has direct knowledge of the algorithm tests. The employee said the change was tested during a four-week pilot in December 2025 on a sample of 2.3 million U.S. users and produced a 44 percent reduction in reshares of flagged posts.

Meta has not publicly announced the change. A company spokesperson did not respond to multiple requests for comment on Saturday and Sunday. The White House Office of Digital Strategy also did not respond to a request for comment.

Coordination with Federal Agencies

The documents outline a weekly coordination process between Meta and federal agencies that has continued despite ongoing litigation challenging the government's role in content moderation. A "weekly sync" call takes place each Thursday at 2 p.m. Eastern Time between Meta policy staff and officials from DHS, the FBI, and HHS, according to the internal memo. The calls are scheduled through April 30, the memo said.

During the calls, agency representatives provide lists of URLs and keyword clusters they want demoted, according to a former content moderator who worked on the trust and safety team until November 2025 and reviewed notes from the meetings. The former moderator, who spoke on condition of anonymity because of a separation agreement, said agency requests were logged in an internal ticketing system called Horizon and assigned priority scores from 1 to 5.

The former moderator said roughly 70 percent of agency requests during the pilot period were categorized as public health related, 20 percent as election related, and 10 percent as foreign influence. The moderator said content from news organizations and elected officials was not supposed to receive higher priority scores than other accounts, but the ticketing system allowed reviewers to apply a "prominent account" tag that could expedite review.

The documents list examples of content that received demotion requests during the December pilot, including posts questioning the safety of a federal vaccine advisory, posts promoting unauthorized election audits, and posts linking to websites identified as Russian or Chinese state media, according to the former content moderator. The moderator said reviewers were instructed to apply the demotion only when the content appeared to violate Meta's existing community standards, though the standard was interpreted broadly during the pilot.

Meta has paid approximately $12 million to an external trust-and-safety vendor to build and maintain the ShieldRank interface, according to an invoice dated Jan. 15 that was reviewed by The Alamo Post. The invoice, sent to Meta's legal department, describes the work as "federal partner integration services" and references a master services agreement signed in August 2025.

Congressional Investigation and Litigation

The planned rollout has drawn scrutiny from congressional Republicans and is expected to be a focus of hearings scheduled for Feb. 10 and Feb. 12 by the House Judiciary Committee, according to three congressional aides briefed on the hearing calendar. The aides said committee staff have requested documents related to Project ShieldRank and the weekly agency coordination calls.

The congressional aides said committee investigators have also obtained copies of emails between Meta policy staff and officials at the Cybersecurity and Infrastructure Security Agency, a DHS component, dating from October 2025 through January 2026. The aides said the emails suggest CISA officials repeatedly asked Meta for data on the geographic distribution of users who shared flagged content.

A lawyer involved in ongoing First Amendment litigation against the federal government said plaintiffs are preparing to amend their complaints to include the ShieldRank system if the rollout proceeds as scheduled. The lawyer, who represents a group of conservative media organizations and individual social media users, said the new documents would strengthen claims that the government is indirectly censoring protected speech by coercing platforms to suppress content.

The lawyer said a federal judge in the Western District of Louisiana is scheduled to hold a hearing on Feb. 18 to consider whether to expand an existing preliminary injunction that limits agency communications with social media companies. The lawyer said Meta employees may be subpoenaed if the documents are not produced voluntarily.

What Happens Next

Meta engineers are scheduled to complete the final code review for ShieldRank on Feb. 6, according to one of the employees familiar with the system. If the review is approved, the phased rollout will begin on Feb. 14 with 10 percent of U.S. users and expand to 100 percent by March 1, the employee said. The employee added that international expansion is not currently planned.

The second employee said Meta's communications team has drafted talking points for a possible blog post, but senior executives have not decided whether to publish the post before or after the rollout. The talking points describe ShieldRank as a "transparent approach to reducing the reach of harmful content while preserving free expression," according to the employee.

The House Judiciary Committee has requested that Meta's chief executive testify at a hearing on Feb. 12, though the company has not confirmed his attendance, according to the congressional aides. The aides said the committee is also considering subpoenas for documents and internal communications related to the Jan. 27 memo.

The Alamo Post will continue to report on the rollout and any public response from Meta or federal agencies.