Bots Don't Sit Still: A Longitudinal Study of Bot Behaviour Change, Temporal Drift, and Feature-Structure Evolution

TL;DR

Promotional Twitter bots exhibit non-stationary, time-evolving behaviour; using a 12-year longitudinal dataset and ten content-based meta-features, the study shows systematic drift and increasing feature interdependencies across generations. The authors demonstrate that newer bot cohorts are more active, media-rich, multilingual, and coordinated in their use of cues, challenging static detection models. They propose a three-phase analysis of feature dependencies, correlation shifts, and evolution patterns to characterize bot adaptation and inform robust, time-sensitive detection systems. The work underscores the need to monitor temporal drift and cohort-specific behaviours to counter evolving bot strategies effectively.

Abstract

Social bots are now deeply embedded in online platforms for promotion, persuasion, and manipulation. Most bot-detection systems still treat behavioural features as static, implicitly assuming bots behave stationarily over time. We test that assumption for promotional Twitter bots, analysing change in both individual behavioural signals and the relationships between them. Using 2,615 promotional bot accounts and 2.8M tweets, we build yearly time series for ten content-based meta-features. Augmented Dickey-Fuller and KPSS tests plus linear trends show all ten are non-stationary: nine increase over time, while language diversity declines slightly. Stratifying by activation generation and account age reveals systematic differences: second-generation bots are most active and link-heavy; short-lived bots show intense, repetitive activity with heavy hashtag/URL use; long-lived bots are less active but more linguistically diverse and use emojis more variably. We then analyse co-occurrence across generations using 18 interpretable binary features spanning actions, topic similarity, URLs, hashtags, sentiment, emojis, and media (153 pairs). Chi-square tests indicate almost all pairs are dependent. Spearman correlations shift in strength and sometimes polarity: many links (e.g. multiple hashtags with media; sentiment with URLs) strengthen, while others flip from weakly positive to weakly or moderately negative. Later generations show more structured combinations of cues. Taken together, these studies provide evidence that promotional social bots adapt over time at both the level of individual meta-features and the level of feature interdependencies, with direct implications for the design and evaluation of bot-detection systems trained on historical behavioural features.

Figures (8)

• Figure 1: Yearly time series (2009--2020) of the ten behavioural meta-features for promotional bots: tweeting, retweeting, replying, URLs, hashtags, duplicated text, sentiment, languages, emojis, and media. Each panel shows the total yearly count across all bots for the corresponding meta-feature.
• Figure 2: Behavioural meta-feature distributions for three generations of promotional bots. Each panel shows the distribution of a meta-feature (posting actions, URLs, sentiment, languages, emojis) for first, second, and third-generation bots.
• Figure 3: Further behavioural meta-feature distributions for three generations, showing hashtags, text, media use, and emoji use, for first, second, and third-generation bots.
• Figure 4: Behavioural meta-feature distributions for three age classes of promotional bots: short-lived (1--4 years), mid-lived (5--8 years), and long-lived (9--12 years). Each panel shows the distribution of a meta-feature for the corresponding age class.
• Figure 5: Analytical framework for characterising behaviour dynamics in promotional Twitter bots. Step 1 computes pairwise dependencies between features. Step 2 quantifies correlations per generation and tracks local/global transitions. Step 3 aggregates these into qualitative evolution patterns.