Here is a summary and humanized version of the report, written in six paragraphs:
The coronavirus pandemic didn’t just create a public health emergency; it created an information emergency that made the health crisis worse. From the very beginning, social media became a battlefield where falsehoods, half-truths, and outright conspiracy theories spread faster than the virus itself. It helps to understand the difference between two related problems: disinformation is the deliberate creation and sharing of false or misleading information, while misinformation is the unintentional sharing of the same kind of content. Both are dangerous, but they feed on each other. Malicious actors intentionally exploit confusion and fear for profit, power, or political gain, while ordinary people, often without realizing it, pass along the harmful content they create. The World Health Organization warned early on about an “infodemic,” and experts quickly realized that public health messages were competing with a tidal wave of bad information. What made it even worse was that some of the most prominent drivers of coronavirus misinformation were not anonymous trolls but politicians, celebrities, and public figures. In fact, research found that top-down misinformation from well-known people made up only a small share of the false claims circulating online, but it accounted for the vast majority of social media engagement. This meant that the people with the largest audiences were often doing the most damage, either by pushing unproven cures, downplaying the threat, or spreading politically charged narratives that made it harder for communities to respond effectively.
As the pandemic enters its next phase, the problem is poised to get worse, not better. During the early months, there was a kind of shared national experience: many states issued stay-at-home orders, and people across the country were facing similar circumstances. That is no longer true. Public health conditions now vary dramatically from place to place, with some communities successfully suppressing the virus, others managing ongoing outbreaks, and still others facing dangerous new surges. This geographic variation is a gift to disinformation producers. They can seize on local events, twist them, and use them to create false narratives that sow division and chaos. The hollowing out of local media over the past two decades makes this even more dangerous. Local newspapers and newsrooms, which once served as a check on misinformation and a source of trusted information, have shrunk or disappeared. That leaves communities informationally vulnerable. When people cannot find reliable local information, they are more likely to turn to social media, where place-based misinformation can flourish. Malicious actors know this, and they are exploiting it. They can misrepresent local conditions, invent fake evidence, and create a false sense of momentum for broader conspiracy narratives. This is not just an abstract worry; it could derail effective public health responses and dramatically distort public perception of the pandemic. Reactive content moderation alone will not be enough to stop this.
The core problem is that social media platforms are designed to maximize engagement, and the very features that make them so addictive also make them ideal vehicles for harmful information. Platforms have long optimized for “frictionless” experiences, meaning they make it as easy as possible for users to share, like, comment, and keep scrolling. Every extra click is seen as a loss. But this design philosophy has a dark side. It incentivizes the spread of outrageous, emotionally charged, and misleading content, because that is what grabs attention. Algorithms amplify the most engaging posts, regardless of whether they are true, and users are given almost no context to help them judge what they are seeing. The report argues that platforms need to fundamentally rethink these design choices and introduce what it calls beneficial friction—intentional obstacles that make people pause and think before sharing. This includes both front-end changes, which users see and feel, and back-end changes, which happen inside the algorithms that decide what content gets promoted. One promising idea is the “virality circuit breaker,” modeled on the circuit breakers used in financial markets. When a stock market goes into freefall, trading is temporarily halted to give investors time to process information and make informed decisions. Platforms could do the same for content that appears to be spreading dangerous misinformation. If a post about the coronavirus is trending and shows signs of being false, the platform could automatically stop amplifying it in newsfeeds and recommendations while human reviewers and fact-checkers take a closer look. This would not prevent people from sharing it, but it would slow the spread long enough to prevent the kind of viral explosion that makes misinformation so damaging.
There are many other practical steps platforms can take behind the scenes. Video autoplay algorithms, for example, have been shown to push users toward increasingly extreme content, and they could easily do the same with dangerous coronavirus misinformation. Platforms should retool autoplay queues so that videos on coronavirus topics lead to authoritative sources rather than conspiracy theories. Accounts that repeatedly share false information should be removed from recommendation algorithms, so they stop showing up as suggested accounts to follow or groups to join. If they keep violating the rules, their followers should be forced to make an active choice: do you really want to keep following this account? These kinds of changes would alter the incentives that currently reward serial misinformation spreaders with growing audiences. On the front end, platforms should develop scan-and-suggest features that analyze draft posts before they are published. If a user is about to share something about the coronavirus that looks like it might be false, the platform could pop up a gentle warning: “Are you sure you’re not spreading false information about COVID-19?” and offer links to credible fact checks. Similar tools are already being tested for bullying and offensive content, and research suggests that asking people to pause and consider the accuracy of a headline reduces the likelihood that they will share it. Direct messaging platforms should also follow WhatsApp’s example by limiting how many times a message can be forwarded at once, which has been shown to dramatically reduce the spread of viral misinformation.
Providing context is just as important as adding friction. Right now, social media strips away the cues that help people make sense of information. A meme, a headline, and a news article all look the same in a feed, regardless of whether they come from a trusted institution or a random account with a hidden agenda. Platforms should embed quality information directly into posts about coronavirus topics. If a post mentions voting, Facebook has already shown that it can automatically attach links to credible voting information; the same approach should be applied to coronavirus posts. Platforms should also give users more information about the people and sources behind the content they see: where the poster is located, how long they have been on the platform, whether they have a history of sharing false information, and whether the website they are linking to has a track record of being debunked. The report also suggests a “truth sandwich” approach, where potentially harmful posts are surrounded by authoritative information from credible sources, so that users are reminded of the facts before and after they encounter the falsehood. Warning labels should carry over when content is shared off-site or embedded elsewhere, rather than disappearing once a post leaves the original platform. And when platforms rely on fact-checkers, journalists, and organizations like Wikipedia to provide this context, they should be fairly compensated for their work. None of these changes will be perfect, and they carry risks, including the possibility that labeling unchecked content might create unnecessary doubt or fatigue. That is why platforms must test these interventions carefully, study their effects, and be transparent about what works and what doesn’t.
Ultimately, this crisis demands more than quick fixes. It requires regulatory action and a structural reckoning with the business models that have allowed misinformation and disinformation to thrive. Social media platforms must be held accountable for the harm they facilitate, and they must be transparent about their algorithms, their enforcement decisions, and the data that researchers need to understand the problem. They should preserve data around COVID-19 content moderation so that independent experts can evaluate their efforts, and they should stop treating content moderation as an afterthought. The workers who review traumatic and dangerous content day after day deserve to be treated as real employees, with fair pay, safe working conditions, and access to mental health support. Platforms should also embrace principles that protect human and civil rights, including the Santa Clara Principles, which call for transparency, notice, and appeals around content moderation decisions. The report acknowledges that some of these recommendations may sound heavy-handed, especially to a generation of designers who have been taught that frictionless experiences are always better. But doing nothing is also a choice, and it is a choice that has already cost lives. The pandemic has exposed the fact that social media platforms are not neutral spaces; they are powerful actors that profit from amplification, and they must take responsibility for the consequences. We are all trying to make sense of a confusing and frightening moment, and platforms have a choice: they can continue to exploit our confusion, or they can change their products to help us find the truth. The public health stakes could not be higher, and the time for urgent, transparent, and structural change is now.