National weather services are the society’s first line of defense against extreme weather events. They are tasked with sounding the alarm about incoming potential natural catastrophes and offering expert advice to the public on which precautions to take during such emergencies. However, since those institutions often serve as well as spokespeople for the scientific evidence on climate change, they are frequent targets of disinformation campaigns and their scientists often receive a significant amount of online harassment.
Those attacks inevitably diminish the national weather services’ credibility and undermine public trust in their alerts, thus endangering lives and property during natural disasters when a part of the public chooses not to heed their advice after being consistently misled into thinking their meteorological agencies are prone to exaggeration or manipulation. As the sustained disinformation campaigns against these institutions often take place at the major digital platforms, it is worth noting that those companies have a legal obligation to address them, at least in the European Union.
A risk and a failure
The substantial loss of confidence in extreme weather alerts fits neatly in the category of “actual or foreseeable negative effects on public security and public health” that the EU Digital Services Act establishes to identify a systemic risk stemming from the use and design of the major digital platforms and search engines. As such, the relevant providers must put in place specific “proportionate and effective mitigation measures” to address such risk, for example by adding information panels or fact-checking labels with contextual information when users see disinformation. However, according to our research, few of the most prominent platforms are currently up to the task.
Fundación Maldita.es has analyzed 169 posts and videos from the last six years that disseminated misinformation about AEMET, the Spanish Meteorological State Agency, in five major digital platforms (X/Twitter, Facebook, Instagram, TikTok and YouTube). While the full results show variation across platforms, only 8% of the videos and posts display any kind of visible label, community note, or information panel that might alert users of manipulative or false content.
A tweet with over 4 million views that falsely accused the agency of lowering the temperature threshold to be able to declare more heat alerts had no visible community note, despite several users proposing notes that debunked it. A YouTube video with 235.000 reproductions said AEMET did not have a working radar in Valencia when flash floods killed over 200 people in October 2024, a lie that the channel also used to ask for economic donations. Many more have seen the agency “admitting” the artificial manipulation of the weather or fabricating evidence of rising temperatures.
Similar viral claims have been debunked in France about Météo-France, in Germany about Deutscher Wetterdienst or in Poland about IMGW-PIB, to name just three examples of national weather services facing disinformation attacks across the European Union. However, the trend extends globally with documented campaigns in different countries, particularly in the United States where the attacks go not only against the weather service but also target disaster relief agencies such as FEMA, with involvement of foreign actors.
A sound strategy to deal with systemic risk under DSA
When designing an effective risk mitigation strategy in this particular field, platforms must keep several important considerations in mind:
As much as they play a crucial role, the meteorological agencies are also government institutions that have to be open to a particularly high level of public scrutiny and criticism, so any risk mitigation strategy cannot focus on removals of content unless it is illegal (i.e. threats, impersonation, fraud, etc.) to not fall into the censorship narrative.
That said, platforms and search engines can and should employ contextual interventions to mitigate disinformation risks by empowering users through fact-checking labels or other similar visual cues at the time they are engaging with false or misleading content, so they can make their own decisions.
As the role of these agencies is most important during extreme weather events, platforms and search engines must ensure their public service information during those periods is highlighted in front of the users that need to see it, for example based on its location, through the deployment of information panels or other mechanisms.
Having those principles in mind, it should be relatively uncomplicated for platforms to build robust risk mitigation strategies that fully align with the legal standards in the DSA, namely that those measures be reasonable, proportionate, effective, and tailored to the specific systemic risks the service presents.
A systematic approach that focuses on showcasing contextual information rather than removing content is reasonable and proportionate, since such a strategy is entirely respectful of the users’ freedom of speech while aptly mitigating the risk of mis/disinformation, as indicated in the particular context of DSA by the European Commission in its guidelines on the mitigation of systemic risks for electoral processes.
Similar strategies have been tested and implemented in all kinds of different services and proved effective across different risk profiles, both reducing the dissemination of dis/misinformation and reducing belief in falsehoods, but also boosting the capacity of dis/misinformation detection by platforms when compared with other singular approaches such as traditional content moderation or crowdsourced fact-checking.
The deployment of information panels to support these agencies’ reach during extreme weather events is also a perfectly reasonable and proportionate risk mitigation intervention, as most platforms and search engines have implemented similar automated displays of third-party information at some point, for example during the COVID-19 pandemic.
To provide users with information about extreme weather events based on their location is not only feasible, it is also common for other prominent information providers that routinely do it and some are required to by law. That is the case of Media service providers under the Spanish Civil Protection Law of 2015 and also of telecom companies that carry public service alerts to cell phones depending on location in many countries. To our knowledge, not one designated very large online platform or search engine displayed information panels, for example, to users in Valencia about AEMET’s alerts regarding the incoming floods that ended up leaving over 200 deaths in October 2024.
In Fundación Maldita.es we know disinformation often harms and even kills, and there are few situations when factual information is more precious than during catastrophes and their immediate aftermath. Weather forecasts and civil protection alerts are literal lifelines in those situations and they need to command the most attention and trust from the public to be effective.
The national weather services themselves are responsible for keeping a public profile that allows them to serve the public best when catastrophes hit, but very large online platforms and search engines have to do their part too by ensuring their policies and the way they are enforced are up to their legal and ethical obligation in this respect.