Eric Siu explains how Google detects AI-generated spam using SentenceBERT to find semantically similar sentences and manual review. Google's system catches the detectable footprint automated AI text leaves behind, so creators should focus on original, data-backed storytelling instead.
Google here, speaking of AI for a second. Like I've said before, there is no way Google lets companies spamming for AI search continue. They will ramp up systems and add manual actions. Google research shows how AI spam can be detected. They're adding in these content defenders. They're using SBERT for identifying AI-generated content. So this is what Google is doing now. Researchers acknowledge that the use of SentenceBERT as a way to identify semantically similar sentences. They cite SentenceBERT to validate a core assumption of their paper that automated AI-generated text leaves a distinct mathematical footprint that can be detected. This system is said to be a highly accurate defense against coordinated generative AI spam. What we've called out even a few months ago, and probably even a few years ago, is that you're better off doing the black hat stuff in a white hat way. If you're going the extra mile with good data-backed content, you're doing good data storytelling, right? That's going to go a lot further than you just stealing someone else's graph and, uh, just reposting it.
