查看文章

Cutting-edge Conversational Artificial Intelligence (CAI) technologies bring ease to human life and the invention of social Artificial Intelligence (AI) bots is one of the ultimate devices in the social media sphere. However, malicious social AI bots lead to societal challenges such as data breaches, information loss, and the proliferation of misinformation. Thus, in this work, significant research has been conducted to address the problem of detecting malicious social AI bots on a microblogging platform such as Twitter. We perform classification through Bidirectional Encoder Representations from Transformer (BERT) embedding-based approach. Utilizing tagging-text features, our preliminary results show the potential of the proposed model to classify tweets as malicious AI bot-generated or human-generated.