Detecting scraper bots through scroll behaviour

Researchers can distinguish between human users and scraper bots by analyzing scroll behavior patterns using burstiness and memory coefficients—metrics that reveal how erratic or uniform scrolling patterns are. Humans typically scroll in bursts with irregular timing intervals when searching for information, while bots tend to scroll more uniformly and predictably. Testing this approach on a machine learning dataset showed promise in differentiating human browsing behavior from AI agents, offering a more reliable detection method than previous techniques that simply identified whether requests came from a real browser.