Prodded by feds, media to police COVID dissent, Big Tech finds consistent standards elusive

Social media COVID censorship in flux with Twitter spam problem, YouTube pullback on masks.

Updated: August 27, 2022 - 5:03pm

Under pressure from federal officials and mainstream media to apply a rigid lens to ever-evolving scientific research, Big Tech companies are flailing in their efforts to treat contested claims consistently.

Twitter blamed an overactive spam filter for wrongly suspending a U.K. doctor who started a project to audit the country's COVID-19 deaths and was previously censored by YouTube for explaining the results of Pfizer's COVID vaccine trial for young children in layman's terms.

The explanation suggests that even as Twitter implements election misinformation rules ahead of the fall midterms, it may be wary of continuing to aggressively police "misleading" COVID tweets in the wake of the CDC's drastic revision of COVID guidance and ongoing publishing of its internal communications related to censorship.

Stanford medical professor Jay Bhattacharya blamed "Twitter interns" for pathologist Clare Craig's permanent suspension over the weekend. A self-described public health physician responded that he too had just been suspended for sharing a study on school mask mandates in Spain.

In her newsletter this week, Craig said she didn't receive notice from Twitter but learned that others who reported her for "hateful conduct" and "abusive behavior" were taking credit for her suspension.

Twitter responded to her appeal by claiming Craig had created "multiple accounts with overlapping uses," evaded a previous suspension and had a habit of "aggressive following." Hours after Craig challenged the factual basis and "started taking advice on legal action," Twitter said her account was wrongly caught in a purge of "automated spam accounts."

It's curious that Twitter didn't credit the goof to "a series of co-ordinated reports" about her abusive and hateful conduct, Craig wrote. "Was the story about spam a way to get them out of admitting that they have a problem with Witch-hunters?"

Twitter's rationale could be seen as a tacit concession of the merits of accusations by its would-be purchaser Elon Musk and former head of security Peiter "Mudge" Zatko.

The billionaire Tesla CEO pulled out of the $44 billion acquisition on the basis of Twitter undercounting spam bots, leading the company to sue him to finish the deal. That trial is set for October

Fired by Twitter in January for allegedly poor performance, Zatko told Congress and federal agencies the company ignored his warnings about chronic security problems that could enable foreign spying, CNN reported. A Twitter manager was recently convicted of spying for Saudi Arabia.

Zatko also alleged Twitter refused to study how many bots were on the platform and had no incentive to do so, which caught Musk's attention. He responded by subpoenaing Zatko.

Tweet URL

The situation got even more confusing Tuesday when Reuters reported a Twitter memo detailed the merger of its "health experience team," which moderates purported misinformation, with its "service team" that tackles spam bots.

Twitter confirmed the "reorganization" to Gizmodo but didn't respond to Just the News queries about the relevance of the team merger to Craig's suspension and Zatko's claims.

Former New York Times journalist Alex Berenson, reinstated to Twitter to settle his breach-of-contract lawsuit, continues to publish Twitter communications from legal discovery that suggest the company suspended him in response to pressure from the feds, the planned defendants in his pending lawsuit

In a mid-March 2021 redacted email thread that includes two addresses marked "senior employee - comms.," one employee relayed a different Twitter team's evaluation of Berenson's account, which "may get more attention moving forward."

While he leverages "individual data points" from sources that are "authoritative" and not, Berenson "avoids making demonstrably false or misleading claims about COVID-19 vaccines," the team wrote, adding that users must be able to "debate and interpret evolving scientific data about the efficacy and safety of the vaccines."

He also found redacted 2021 emails from CNN's Oliver Darcy and Axios's Ashley Gold asking Twitter media relations why it was "boosting anti-vax" Berenson and letting him spread "a lot" of "Covid vaccine misinformation." Darcy's request came days before Berenson's suspension.

YouTube quietly removed masks entirely from its COVID misinformation policy, a change apparently first widely publicized by podcaster Tim Pool this week. Archives of the page show the language was removed between May 9 and May 17. The video sharing giant didn't respond to queries about the timing and basis for the removal.

The page formerly said YouTube would remove content that claims "wearing a mask is dangerous or causes negative physical health effects" or that "masks do not play a role in preventing the contraction or transmission of COVID-19," and possibly sanction users. 

Tweet URL

Conservative media personality Dan Bongino claimed YouTube suspended his channel in January "for quoting Dr. Fauci saying masks don’t work (one of his countless flip-flops)." Last summer, Sen. Rand Paul (R-Ky.) said YouTube suspended his channel for challenging mask efficacy. 

Facebook's external fact-checker for COVID misinformation, whose judgments are used to throttle traffic and sanction users, got an unflattering review from University of California San Francisco epidemiologists in the academic journal Digital Health.

Alyson Haslam and Vinay Prasad said they searched articles on Health Feedback for the "names of reviewers and other quoted individuals cited in the article[s] and their professional titles," then searched Twitter for their associated accounts and size of their Twitter followings.

The median follower count for reviewers was 10 times larger (10,000) since January 2020 than before the pandemic started drawing international attention (1,012). The difference was staggering for quoted individuals: 153,050 median followers since January 2020, 314 before. 

Haslam and Prasad found at least one reviewer since January 2020 who had "already been critical of the topic on Twitter" before reviewing the article.

It is "concerning" that HealthFeedback.org is selecting reviewers with such large followings, who then in turn quote individuals with even larger followings, Haslam and Prasad wrote. "Further, reviewers appear to be selected based upon a priori viewpoints that reach a large audience," they said, calling for more transparency in how articles and reviewers are chosen.

Tweet URL

Just the News Spotlight