How Facebook Fails to Safeguard Hungarian Elections

Illegal ads, fake profiles, and AI-generated content could unfairly affect the election - writes Lili Rutai.

Despite an EU-wide ban on political advertising, organisations affiliated with Hungary’s governing Fidesz party continue to run paid political ads on Facebook, reaching millions of users just months before a decisive general election in April. At the same time, thousands of fake profiles help boost engagement for government politicians, churning out deepfake videos and tricking the algorithm. The platform’s failure to filter these illegal ads could influence the outcome of the vote.

Facebook is the most widely used social media platform in Hungary: with nearly 7 million users, more than 70% of the population is active on the platform. While younger generations often opt for Instagram or TikTok instead, Facebook remains a frequently visited platform and a trusted news source, especially for the elderly, who make up most of Fidesz’s voters and may not be well-versed in deepfakes that are shared on the platform - even on the official page of Prime Minister Viktor Orbán.

At least ten pages and organisations linked to Fidesz are still promoting content that is clearly political in nature, analysis of Meta’s Ad Library shows. Several of their posts are misclassified or use AI-generated or deepfake imagery that platform moderation systems appear unlikely to detect. While some ads were eventually removed, dozens ran long enough to reach millions of users.

Undecided voters still make up at least a quarter of the electorate. Less than three months before the election, which side will persuade them could determine the future of Hungarian democracy. Facebook, and other social media platforms play a key part in this.

The anatomy of an illegal ad