On November 10, 2026, the European Commission published a draft update of its election guidelines under the Digital Services Act and opened a consultation on it. The guidelines tell the largest platforms and search engines, the ones the law calls very large, how to assess and reduce risks to elections. The first version dates from April 2024. The Commission says the update responds to generative AI and to "increasingly sophisticated manipulation risks." Comments are open until November 23, 2026.
This is a draft, and it applies to platforms, not to parties, unions, or NGOs. But several of its changes affect how an organization's posts travel during a campaign, and how fast it gets help when someone attacks its candidates.
What the draft asks platforms to do
The Commission lists five main changes. In its own summary:
- Foreign interference: platforms should detect coordinated inauthentic behavior early and dismantle the networks behind it.
- Recommender systems: platforms should "spot and assess suspicious trends before they go viral," test their algorithms for manipulation before every national election, and keep a record of changes to their recommender systems in the six months before a vote.
- AI-generated content: deepfakes should be clearly labeled and keep their label when shared, and illegal deepfakes of candidates should come down fast.
- Candidates: dedicated channels to platforms, and swift handling of threats, harassment, and impersonation, "applied neutrally to all candidates."
- Accountability: platforms should report to the Commission and national authorities within 30 days of an election, and publish a public version.
The draft also says that "nothing in these guidelines suggests the removal, demotion or labelling of lawful content on account of the political opinion it expresses." It targets manipulation of the services, not opinions.
Why this matters for your team
Read paragraph 85 closely. When a platform sees a sudden spike in engagement around a topic, hashtag, or account, the draft asks it to check for coordination and inauthentic accounts. Where it finds "objective signals of coordination or inauthenticity," the related content and accounts "should be temporarily excluded from recommendations" until the check is done. The draft says this should be lifted "without delay" if no breach is confirmed, and that it should not be applied because of political views.
That is a protection against the networks that flood a debate. It also raises a practical question for any organization that asks many supporters to post at once. A hundred real people posting the same pasted sentence in the same ten minutes can look a lot like the pattern platforms are told to catch.
The candidate measures point the other way. If platforms build dedicated reporting channels for candidates and their authorized representatives, the team that already knows who its contact is, and has its evidence ready, is better placed to get a fast answer.
Four habits for the months before a vote
- Ask for posts in your supporters' own words. Give people the facts, the reason, and the link, and let them write the post themselves. Posts written by real people tend to vary, and varied wording is less likely to resemble the copy-paste pattern the draft describes.
- Keep a dated incident log. For each attack on your candidates or spokespeople, note the time, the links, screenshots, and what you reported to whom. You'll need it for a platform's candidate channel, and the draft also asks platforms to work with civil society organizations and share risk assessment results with them.
- Sort out reporting before the campaign starts. The draft expects platforms to onboard candidates "sufficiently in advance of the electoral period." Find out who on your team is the authorized contact for each candidate, and report through those channels. Asking supporters to mass-report an account is not a substitute, and it can look like coordinated abuse itself.
- Label your own AI images and video. If your team uses AI to make a campaign graphic, label it before it goes out. The draft asks platforms to keep labels attached when content is shared, and an unlabeled synthetic image is an easy target for the other side.
If your organization faced attacks in a recent election, the consultation is a direct way to say what helped. The Commission invites "all stakeholders with relevant experience and expertise" to respond by November 23.
Where Ostracon fits
Ostracon is built around the first habit. A coordinator gets a briefing that recommends whether to ignore, mirror, or push, with ignore as the default when the post is bait, and a person on the team decides. Missions go to consenting, unpaid supporters, who post in their own words by default. Every mission stays in a permanent archive your team can show when someone asks what you did. Ostracon doesn't send mass-reporting missions.
Sources
- Commission invites views on draft revised election guidelines under Digital Services Act, European Commission press release, November 10, 2026
- Targeted consultation on the draft revised election guidelines under the Digital Services Act (DSA), European Commission, November 10, 2026
- Draft revised election guidelines under the Digital Services Act, European Commission draft for consultation, November 10, 2026
- Guidelines for providers of VLOPs and VLOSEs on the mitigation of systemic risks for electoral processes, European Commission, April 2024