Generative AI is making online propaganda faster and harder to spot

In short: Researchers and analysts say generative AI is enabling a new style of propaganda that is cheaper to produce, easier to tailor, and harder to detect.

What's going on

Generative AI means tools that can create new text, images, audio, or video on demand, like a machine that can write endless variations of the same message. In propaganda, that can translate into thousands of posts, comments, emails, or articles made in minutes, each worded a little differently. Studies comparing human propaganda to AI written propaganda have found the AI versions can be similarly persuasive, especially when a person lightly edits the output.

Another change is targeting. Instead of one message blasted to everyone, AI can help tailor messages to specific groups, or even to individuals, based on what they seem to care about. That can include chatbots that hold long conversations and adjust in real time, and robocalls that use fake audio made to sound like a trusted person.

Video and audio fakes, often called deepfakes (real looking media made by AI), are also improving. Examples cited in recent reporting include AI generated news anchors used to push pro government narratives, and edited clips that make political leaders appear to say things they never said.

What to watch

A key risk is that these campaigns can look like normal public chatter, since AI can avoid obvious copy and paste patterns. Another worry is a broader “fog” effect, where people start doubting real evidence because fakes are so common. Expect more debate over labeling AI made political content, and stricter rules on fake accounts and election related deepfakes.

Source: NYTimes