Editor’s Note: MRC got the idea for its research utilizing xAI Grok from a query posted by an X account @The1Parzival. The published study omitted a hat tip.
Either Jack Dorsey's ghost still haunts X's algorithm or, alternatively, a radical remnant continues to work its dark arts within the platform’s ranks to undermine Elon Musk's vision of a free-speech platform divorced from its history of election interference.
A new study by MRC Free Speech America researchers found that the X algorithm has actively pushed Vice President Kamala Harris’s account, granting her a broader reach than former President Donald Trump’s potential GOP running mates.
The study, performed using X’s open-source algorithm and its AI chatbot Grok, comes immediately after another MRC report revealed that X was de-amplifying the accounts of Republican members of Congress while simultaneously boosting radical Democrats.
This favoritism is also evident in the 2024 veepstakes, with X promoting Harris significantly more than North Dakota Gov. Doug Burgum and Sens. JD Vance (R-OH) and Marco Rubio (R-FL), who are all on Trump’s VP shortlist.
Here are some of the study’s key findings
- X Turns Blind Eye to Harris’s Violations. Harris consistently violates the algorithm’s criteria for censorship but is still shielded by the platform, likely helping her fundraise and recruit voters online.
- Community-Noted and Fact-Checked Favor. Harris has used X to spread demonstrably falsehoods and vitriolic attacks against Republicans, yet continues to receive high credibility ratings from X’s algorithms.
- X Punishes Trump’s Potential Running Mates. The algorithm limits the reach of all of the contenders for Trump’s VP pick. However, according to X’s metrics, Burgum and Vance face significantly harsher restrictions, indicating their reach could be largely limited.
- The Bias Exposes an Internal Revolt Against Musk. While Musk has pledged to protect freedom of speech on his platform, the biased algorithms appear to hint at the existence of partisan rank-and-file X employees running cover for Harris.
X’s Algorithms Explained: On Monday, the Media Research Center released a report using X’s AI Grok to reveal how X’s algorithms use four core metrics (ranging from 0 to 100) to suppress or promote accounts.
According to Grok, these factors include “Mass Appeal” (diversity of followers), “Reputation” (purported reliability), “Toxicity” (potentially offensive content or perceived harmfulness) and “Follower” (the account’s follower retention). Each category is weighted differently in how it impacts the overall visibility score, which is then used to determine accounts’ reach.
Notably, as explained in MRC’s previous report on X, the reputation category largely determines “the visibility and likelihood of a user's content being recommended to others.”
While Harris enjoys a favorable reputation score of 72, Rubio has a score of 70, and both Burgum and Vance have scores of 50. This means Burgum’s and Vance’s content is rarely promoted by X.
X Keeping Score: Harris boasts a visibility score of 79.53 out of 100, enjoying high scores in mass appeal (85), reputation (72), toxicity (25) and follow (90). This means X has actively boosted Harris’s tweets amid her shortcomings as vice president – scathingly low polls and mounting scandals.
Harris’s highly favorable score is strikingly different from those granted to Trump’s potential running mates. Specifically, Rubio received a visibility score of 70.5, while Burgum scored 62.87 and Vance scored 58.12.
Such negative numbers showcase Harris’s undue advantage over Republican VP hopefuls with a 9-point lead over Rubio—and a 16-point and 21.4-point lead over Burgum and Vance, respectively.
But there’s one kicker: According to X’s censorship metrics, Harris should be de-boosted due to her record of posting demonstrably false claims and vitriolic content against “MAGA” Republicans.
Harris has defended Biden’s cognitive decline, falsely accused Trump of wanting to cut Social Security and Medicare and touted a non-existent immigration record. She also misleadingly claimed that the Biden administration’s policies have reduced inflation, which fact-checkers have disputed.
Harris was fact-checked after boasting of the creation of “more than 13 million new jobs” thanks to “Bidenonics,” On Aug. 3, 2023. X’s Community Note correctly slapped Harris’s post with a context banner, clarifying that the so-called newly created jobs were mere bounceback positions following the COVID-19 pandemic.
On June 28, 2024, Harris refuted fact-based evidence of Biden’s deteriorated physique and forced gait by suggesting Biden is sharp when the cameras are off. Harris’s assertions directly contradict descriptions from Biden’s allies that claim the president has appeared diminished and confused in private meetings.
In June 2020, Harris prompted a bail fundraiser for individuals arrested and charged with violent crimes in Minnesota. “If you’re able to, chip in now to the @MNFreedomFund to help post bail for those protesting on the ground in Minnesota,” she wrote.
Among the individuals bailed by the Minnesota fund were defendants charged with murder, violent felonies and even child rape, according to Fox 9 Minnesota.
In several posts, Harris has also widely accused Trump of wanting to cut Social Security and Medicare—a claim that even her lefty friends at The Washington Post have fact-checked. “Trump has explicitly ruled out any changes in Social Security,” wrote The Post’s Glen Kessler on Wednesday.
Harris has also claimed to be working to “fix our broken immigration system” despite record-high border encounters under the Biden-Harris administration. This has culminated in nearly 10 million encounters nationwide, according to Homeland Security data cited by House Republicans.
Moreover, Harris has falsely claimed the Biden-Harris administration’s policies have helped “reduce inflation.” This claim, according to Democrat-friendly FactChek.org, is misleading.
The fact checker cited economists who said Biden’s policies played a “modest role” in worsening inflation, which reached an average monthly rate of 5.4%, compared to Trump’s 1.9% monthly average.
Most recently, Harris was stung by Community Notes, X’s fact-checking tool, after falsely accusing Trump of backing an “abortion ban nationwide.” Harris’s post was flagged by a fact-check that read, “President Trump has repeatedly said he will not sign a national abortion ban.”
In contrast, Rubio and Burgum have not received community notes. Vance’s only note was related to his past criticism of Trump, something not uncommon for politicians. For instance, Harris rebuked Biden for opposing mandatory busing for desegregated schools.
Implications: X’s giving Harris an algorithmic advantage suggests that the platform is boosting a widely unpopular candidate against two important backdrops: the Republican National Committee's convention and Harris’s potential elevation to the presidency.
Biden’s cognitive decline and polls suggesting he could be a one-term president are just adding fuel to the fire of Democrats’ concerns about whether Biden should be the Democratic Party’s 2024 presidential nominee. Harris herself has struggled with low polls, with many showing she is more unpopular than her boss. However, Harris benefits from X’s algorithms and longstanding support from Big Tech giants, including Google.
In May, Musk pledged to address the issue of suppressed content when pressed by Sen. Mike Lee (R-UT) on X. “Well, neither conservative [sic] nor progressives should be throttled. The point is to have an even playing field,” Musk responded. “I will investigate.”
It isn’t immediately clear whether Musk has directed his employees to address this issue or if they have ignored his commands. This situation suggests that the ghost of the Jack Dorsey-owned Twitter 1.0 is still affecting the platform and targeting Republicans, despite Musk's pledge to change this.
MRC Free Speech America Intern Christian Baldwin contributed to this report.
METHODOLOGY: Between June 7 - June 25, MRC researchers gave Grok specific prompts to ascertain “visibility” scores for Vice President Kamala Harris and three of Trump’s rumored vice presidential candidates.
The questions were specifically tailored so Grok would only report the essential criteria that would determine whether or not a specific account on X is pushed/recommended to other users on the platform (“Mass Appeal,” “Reputation,” “Toxicity,” and “Follow”). Grok calculated numerical value scores between 1 and 100 for each individual. MRC researchers then calculated the average scores for all members of Congress across all four criteria used by the algorithm to determine X’s “visibility” scoring process.
To determine how X weighted each of the four categories in determining a “visibility” score, MRC researchers prompted Grok with eight tests between July 1 and July 2 to assess how each of the metrics were weighted to uncover which was considered most important when determining whether or not an account would be boosted.
Grok repeatedly provided information with minor variations on how each category was weighted to determine a congressional member’s final “visibility” score.
MRC researchers then took the average weighted percentages across the eight tests to deduce a mathematical formula. When the averages were computed, the categories were weighted as follows: Mass Appeal (38 percent), Reputation (30 percent), Toxicity (21 percent), and Follow (11 percent). “Mass Appeal” and “Reputation” scores were consistently in the top two spots in measures of importance across the eight tests.
Conservatives are under attack. Contact your representatives and demand that Big Tech be held to account to mirror the First Amendment while providing transparency, clarity on hate speech and equal footing for conservatives. If you have been censored, contact us using CensorTrack’s contact form, and help us hold Big Tech accountable.
