Germany's AfD built an AI tool to mass-produce “rage bait” using chatbots like Claude
Anti-immigration and homophobic content warning.
Infectious AI ideologies. Image by Cybernews.
- Correctiv exposed AfD’s Alternita platform, which uses ChatGPT, Gemini, and Claude to generate political social media posts.
- Undercover tests found the tool created anti-immigration, racist, and anti-LGBT posts, including calls for involuntary remigration.
- The system reportedly accepted neo-Nazi source material and generated new content without blocking it, raising moderation concerns.
- Unlabeled AI-generated campaign posts could mislead voters as AfD remains Germany’s largest opposition party before 2026 regional elections.
Key Takeaways by nexos.ai, reviewed by Cybernews staff.
The far-right German party Alternative für Deutschland’s (AfD) new AI platform, Alternita, has been exposed in an undercover investigation, as having used homophobic and racist propaganda to help proliferate its message across social media.
The investigation was carried out by Correctiv, a German investigative newsroom that obtained access to the platform by posing as an AfD party member.
The AfD, or Alternative for Germany, is currently one of the country's strongest polling parties. It uses OpenAI's ChatGPT, Google Gemini, and Anthropic Claude as its AI backbone.
Correctiv says the platform integrates leading commercial AI models rather than using a proprietary model, and is designed to create social media posts "in five minutes, with your positions, in your style, your branding."
The platform is marketed as a way for party members to rapidly generate campaign content while maintaining consistent messaging.
AI-generated posts calling for "consequential and involuntary remigration" after uploading an article were found in the Correctiv investigation, with the undercover reporters prompting the AI to produce more provocative language based on an immigration-related news story.
Reporters also uploaded content from a neo-Nazi publication to test the system, with the software reportedly accepting the material and generating new posts without blocking the source.
When Cybernews investigated an aspect reported by The Irish Times, we found AI-generated content on the X channel of fictional politician Karl Ranseier, including a video of immigrants being sent home on a plane and synthetic crowds chanting "send them back” over dubious rave music.
People across Europe have had enough of politicians who govern against them. “Send them back” is now echoing across the continent — and Keir Starmer will be only the first domino to fall.
undefined Bundesremigrationsminister Karl Ranseier (@Ranseier_Karl_) June 22, 2026
Music: Continental Pulse - Send them back pic.twitter.com/kda7U6aVkn
The platform also generated anti-LGBT messaging alongside AI-created imagery. One output included a post warning of a "rainbow dictatorship" accompanied by an AI-generated image of a burning rainbow flag.
The investigation suggests not all AI-generated content is labeled as AI, raising transparency concerns if voters cannot distinguish AI-written political messaging from human-created posts.
One of the AI-generated posts produced during Correctiv's test of the platform even called for "an immediate halt to naturalizations... to prevent ethnic voting blocs and protect our homeland" in a strong tirade on immigration.
The far-right party is the largest opposition party in Germany and is heavily favored to gain power in regions of eastern Germany in the upcoming September 2026 elections.
