Landtagswahl Sachsen-Anhalt 2026: Desinformation auf X

Why this data matters

State elections are increasingly fought on platforms before they're fought at the ballot box. Understanding a race means understanding how it's framed online — which claims spread, which narratives repeat, and which ones don't survive contact with the facts. That's the gap this dataset fills: 7,612 X posts tied to the Saxony-Anhalt election, coded and checked rather than skimmed.

This kind of systematic review — separating genuine political debate from claims that misrepresent events — is close to what our own AI text detection work is built for: telling manufactured or manipulated content apart from the real thing at scale.

Methodology, briefly

The conservative "pro-party" rule matters. Simply counting mentions of "AfD" would be misleading, since a large share of the dataset criticizes or neutrally reports on the party rather than endorsing it.

The party split: AfD vs. everyone else

That ratio applies only to those 215 endorsement posts, not the full 7,612-post dataset. The AfD is mentioned in an estimated 3,000-plus posts overall, but a mention isn't support. The dataset also contains a meaningful anti-AfD and tactical-voting current, plus a distinct BSW peace/anti-escalation thread.

Six narratives carrying the conversation

Topic frequency across the full dataset

Categories overlap — a post about Ukraine can also fall under sanctions. What stands out is the sheer weight of foreign-policy topics relative to the dataset's size. Russia, Ukraine, war/peace, and the EU aren't background noise here; they're repeatedly tied directly to the state race.

How Russia and Ukraine show up in pro-AfD posts

Inside the 149 strictly coded pro-AfD posts, one narrative package repeats:

Germany and Brussels are escalating the war → Ukraine policy is hurting Germans → Russia is being unfairly demonized → the AfD is the peace party → voting AfD is therefore both a foreign-policy and domestic correction.

The recurring building blocks: framing Russia as the victim of Western escalation (terms like "Russophobia," Nord Stream cited as proof of Western double standards, sanctions called "dangerous" or "irrational"), and an anti-EU frame that positions Brussels as part of a detached, "globalist" establishment.

One clear example is a post from AfD figure Björn Höcke calling the federal government and its EU partners "warmongers" and drawing the conclusion that Saxony-Anhalt should vote AfD as a result. The AfD's stated position on closer ties with Moscow and ending support for Ukraine also lines up with independent reporting, including from Reuters.

Worth being precise about: the data shows a coherent, ideologically organized narrative ecosystem that overlaps heavily with AfD campaign messaging and shares features with known Russian influence narratives. It does not, on its own, prove technical coordination by a Russian state apparatus. Near-identical, repeated posts are mostly explained by amplifier accounts, quote-retweets, and circulating source material — not necessarily bot networks.

The Leipzig/Halle drone incident: the campaign's sharpest disinformation signal

The most striking disinformation vector in the dataset centers on a drone incident near Leipzig/Halle shortly before the election. A recurring claim holds that the attribution to Russia was fabricated — a "false flag" staged to block an AfD win, specifically Ulrich Siegmund becoming state premier.

Timing matters here. In early-to-mid August 2026, the federal government was still cautious about attributing the incident; the investigation was ongoing and no attribution had been made. Some skeptical posts from that window weren't necessarily false — they reflected genuine uncertainty at the time.

That changed on September 1–2, 2026. According to reporting from NDR, WDR, and Süddeutsche Zeitung, German authorities now believe the Russian government was behind the attempted attack; suspects with ties to Russia have been identified, and Federal Interior Minister Dobrindt pointed to evidence of action on behalf of Russian state entities.

Posts that keep presenting the "fabricated attribution" claim as settled fact after that point count as misleading and outdated, not legitimate uncertainty. It's the kind of claim where verifying the underlying source and timeline — rather than trusting the framing — is what separates a fair read from a manufactured one, which is exactly the gap tools like our AI scam and deepfake detector are built to close for individual claims and media.

Claims we checked

Not every contested claim here is disinformation. The table deliberately separates demonstrably false claims from unsupported ones, outdated-but-once-reasonable ones, and ordinary political opinion.

What this analysis does not show

The boundaries matter as much as the findings. This dataset does not prove a Kremlin-directed bot network — repetition and quote-retweets point to organic amplification through multiplier accounts, not necessarily state control. And 69.3% doesn't mean "69% of all posts are pro-AfD" — that figure applies strictly to the 215 posts with an explicit endorsement, not the full 7,612. Finally, not every Russia- or EU-critical post is disinformation; there's a legitimate policy debate happening here too, which is why the claims table above separates false claims from unsupported ones and ordinary opinion.

Frequently asked questions