There is a concern among reputable news media that their content (both content published by them and posted by them, as well as posted by readers) is vulnerable to arbitrary content moderation by social media companies. This debate takes place against the backdrop of two major EU media and digital policy frameworks. The Audiovisual Media Services Directive (AVMSD) sets common EU rules for broadcasters, streaming services, and video-sharing platforms. The European Media Freedom Act (EMFA), which entered into force in 2024, introduces new safeguards aimed at protecting media freedom and media pluralism. Discussions around the future of the AVMSD have raised the possibility of extending or replicating elements of EMFA’s Article 18, which grants certain procedural protections to media service providers when platforms seek to restrict or remove their content. It is in this context that proposals for a “media exemption” have emerged.
In an environment where revenue is closely tied to “clicks”, this translates to legitimate concerns from reputable news media about the weak position this puts them in vis-a-vis the increasingly political social media companies. A media exemption, where their content could not be limited in this way, is often touted as a solution. What has been consistently missing is data on whether the concern, however legitimate, is actually a problem in the real world.
Concerns about EMFA
EMFA Article 18 allows media service providers to obtain privileged treatment through a self–declaration to very large online platforms. This means in practice that content posted by self-declared media being exempted from quick removal and those media having access to an appeals process that is not available to regular users. As noted during the EMFA negotiations, this amounts to a media exemption from the normal rules of content moderation.
Instead, the platform seeking to implement content moderation is obliged to:
- Provide prior notification within a notice period of 24 hours together with a justification of their planned actions, if the content violates its terms of use;
- Conduct good faith negotiations with the service provider during this time with a view to reaching an amicable resolution;
- Deal with complaints of these providers first under its internal complaints mechanism.
This approach raises some clear problems:
- Most urgently, in an environment where “clicks” equate to financial revenue, even a partial exemption from content moderation creates perverse incentives – if outlets have impunity to publish awful-but-lawful content designed to gather clicks, this will inevitably be abused.
- There is little or nothing to prevent illegitimate news outlets self-declaring to social media companies that, in turn, have no incentive to verify the authenticity of claims.
As we see with the large-scale proliferation of scam ads on social media, the social media companies already appear to see little incentive to address malicious content on their services. Consequently, it seems unlikely that social media companies seem unlikely to want to take meaningful action against abusive self-declarations.
Why a “media exemption” in the AVMSD should be avoided
- Inadequate regulatory oversight: privileged treatment would at least need to be accompanied by a mechanism whereby a self-declaration would automatically also be copied to the national media regulator, for at least minimal oversight and accountability. The Commission rejected even that minimal safeguard in EMFA, so is unlikely to support it in any future instrument.
- It is impossible to define the “good” media that deserve to be protected: Bad-faith actors will exploit any exemption that makes it difficult to remove their click-bait, disinformation or other low-quality content. The behaviour of the Dutch Public Broadcaster, as illustrated by Ongehoord Nederland also shows the difficulties of drawing the line between “good” and “bad”.
- No “due prominence” turned into an de facto whitelist: Article 7a of AVMSD rule needs remain confined strictly to services which perform a public interest mandate. Expanding its scope to include a more generic class of “trusted” media opens the category also to media with malicious intentions.
- Overlap with the DSA: the Digital Services Act gives social media users extensive new rights to appeal against, and gain more information about, content moderation decisions that they have been subjected to. In two years, fifty million decisions were reversed under these new rules. This makes it even more unlikely that a specific ex ante protection mechanism is needed for legitimate media outlets.