AI-generated content is getting harder to spot

Artificial intelligence is making it increasingly difficult to tell real social media content from fabricated material, ORF reports. Algorithms learn from users’ behaviour and increasingly use AI to select what appears in their feeds.

Jana Lasser, a professor of data analysis, said the systems are designed to keep users engaged by showing them content likely to hold their attention. Stepping back to reflect on what they are consuming can be difficult, she said.

Many social media profiles also feature AI-generated people and content. For the past two years, the ORF defacto newsroom in Vienna has checked videos, images and information for manipulation. Eva Wackenreuther of the newsroom said it uses tools to identify AI material, but also applies journalistic standards rather than relying on software alone.

Verification is especially important during breaking news, when events may be unclear and emotions are running high, Wackenreuther said. Such moments can be used to spread false information. She said checking content before it is broadcast takes time and staff, but is essential to maintaining trust.

Disinformation can also generate revenue. Lasser said recent findings indicated that Meta earns 10 percent of its advertising revenue from illegal content, adding that the company knows about this but has little incentive to act because it makes money from it. She said people are fed up with the growing flood of AI content on social media.